Challenges To Democracy in Latin America

Chhaalllleennggeess ttoo Deemmooccrraaccyy iinn
Laattiinn Ammeerriiccaa aanndd tthhee Caarriibbbbeeaann::

Evviiddeennccee ffrroomm tthhee AmmeerriiccaassBaarroommeetteerr 22000066--0077

Mitchell A. Seligson, Ph.D., Editor

The publication of this report was made possible thanks to the
support of the Democracy and Governance program of United
States Agency for International Development. The opinions
expressed herein do not necessarily reflect the views of the
United States Agency for International Development.
March, 2008
Challenges to Democracy in Latin America and the Caribbean:
Evidence from the AmericasBarometer 2006-07
CONTENT
Preface................................................................................................................................. v
Foreword........................................................................................................................... vii
The AmericasBarometer, 2006-2007: Background to the Study ............................................................ vii
Acknowledgements ................................................................................................................................... x
Overview ..........................................................................................................................xiii
Part I. Challenges to the Institutional Infrastructure of Democracy............................ 1
I. The Rise of Populism and the Left: Challenge to Democratic Consolidation? .... 3
Mitchell A. Seligson
Abstract ..................................................................................................................................................... 3
The Political Orientation of Latin Americans is Important to Democracy ............................................... 5
Political Orientation Matters: People Vote their Beliefs..........................................................6
Leftists are Less Likely to Favor some Core Democracy Values ..........................................10
Latin Americans are not Ready to Jettison Democratic Institutions ....................................................... 13
Conclusions: Political Institutions and Values Remain Important.......................................................... 16
References ............................................................................................................................................... 18
II. Illiberal Democracy and Normative Democracy: How is Democracy Defined in

the Americas?............................................................................................................ 21
Julio F. Carrin
Abstract ................................................................................................................................................... 21
Mass Conceptualizations of Democracy ................................................................................................. 23
How Consistent is the Support for Democracy? ....................................................................30
Who Is More Likely to Embrace Liberal Democracy? ........................................................................... 37
Concluding Remarks............................................................................................................................... 45
References ............................................................................................................................................... 47
III. Throw them All Out? Attitudes towards Political Parties in the Americas.... 53
Mara Fernanda Boidi
Abstract ................................................................................................................................................... 53
Weak Citizen-party Linkages in the Americas ....................................................................................... 54
Why Trust in Political Parties Matters .................................................................................................... 57
Political Parties and Democratic Consolidation in the Region ............................................................... 67
References ............................................................................................................................................... 71
IV. Economic Performance and Support for the System: Economic Challenges for
Latin American Democracies .................................................................................. 73
Vivian Schwarz-Blum
Abstract ................................................................................................................................................... 73
Development, Economic Growth and Support for the System ............................................................... 84
Conclusions............................................................................................................................................. 86
References ............................................................................................................................................... 88
i
Challenges to Democracy in Latin America and the Caribbean: .
V. Decentralize or Centralize? Challenges for Reform of the State and Democracy

in Latin America and the Caribbean...................................................................... 89
Daniel Montalvo
Abstract ................................................................................................................................................... 89
Decentralization, Centralization and Intergovernmental Transfers ........................................................ 90
The State of Decentralization in Latin America and the Caribbean........................................................ 92
Public Opinion: Political Trust among Institutions and Support for Decentralization ........................... 94
Comparative Level of Municipal Political Trust versus Other Democratic Institutions......................... 97
Political Trust in Municipal Government per Country ..........................................................98
Political Trust in the Municipal Government Compared with Trust in the National
Government, per Country.......................................................................................................99
Variables that Determine Political Trust in Municipalities in Latin America and the
Caribbean..............................................................................................................................102
Centralize or decentralize? .................................................................................................................... 104
Conclusion ............................................................................................................................................ 109
References ............................................................................................................................................. 111
VI. Is the Vote for the left a Risk or Opportunity for Democracy in Latin America?.
................................................................................................................................ 119
Rosario Queirolo
Abstract ................................................................................................................................................. 119
First Argument: Voting Left as a Backlash Against the Washington Consensus ................................. 121
Theories, Hypotheses and Country Cases ............................................................................................. 123
Economic voting theory .......................................................................................................123
Cleavages Created by Political Process................................................................................125
The Cases .............................................................................................................................126
The Model and Results.......................................................................................................................... 128
Results................................................................................................................................................... 130
Second Argument: Voting Left as a Risk for Latin American Democracy ......................................... 133
Results ..................................................................................................................................134
Conclusion ............................................................................................................................................ 139
References ............................................................................................................................................. 141
Part II. Challenges to Democracy from Civil Society................................................. 145

VII. Social Trust, Economic Inequality, and Democracy in the Americas ............... 147
Abby B. Crdova Guilln
Abstract ................................................................................................................................................. 147
A Brief Review of the Literature .......................................................................................................... 150
Social Capital in Latin America in Comparative Perspective..............................................152
A Weak Link: Social Trust and Civic Participation.............................................................155
Explaining the Social Trust gap in the Americas: Do Individual Characteristics Matter? ................... 159
Socio-economic Characteristics ...........................................................................................159
Demographic Characteristics ...............................................................................................160
Personal Experience with Crime ..........................................................................................162
Economic Inequality and Social Trust: A Multi-level Analysis ........................................................... 164
The Importance of Social Trust for Latin American Democracies ....................................................... 167
Conclusion ............................................................................................................................................ 170
ii
References ............................................................................................................................................. 171
VIII. National Identity and Ethnic Minorities in the Americas................................. 177

Daniel Moreno
Abstract ................................................................................................................................................. 177
The National Political Community ....................................................................................................... 178
National Belonging: Is it Necessary for Democracy? ........................................................................... 178
The Strength of the National Political Community in the Countries Studied ....................................... 180
National Belonging and Ethnic Identity. Questioning Assumptions .................................................... 184
Combining Individual Factors and Contextual Factors......................................................................... 189
Conclusions........................................................................................................................................... 191
References ............................................................................................................................................. 193
IX. Challenges of Tolerance in the Americas ............................................................. 195

Diana Orcs
Abstract ................................................................................................................................................. 195
Political Tolerance ................................................................................................................................ 196
Why Does Political Tolerance Differ Across Countries? ....................................................198
Which Individual-level Factors Influence Differences in Political Tolerance? ...................199
North America and the Caribbean........................................................................................200
Central America ...................................................................................................................201
South America......................................................................................................................203
Methods and Results ............................................................................................................................. 205
Social Tolerance.................................................................................................................................... 208
Methods and Results ............................................................................................................................. 211
Conclusion ............................................................................................................................................ 212
References ............................................................................................................................................. 214
Part III. Challenges to the Rule of Law ....................................................................... 217

X. The Impact of Violent Crime on the Political Culture of Latin America: The
Special Case of Central America........................................................................... 219
Jos Miguel Cruz
Abstract ................................................................................................................................................. 219
Conditions for Democracy, Political Culture, Legitimacy and Violence.............................................. 220
Criminal Violence in Latin America..................................................................................................... 223
Methodological Aspects........................................................................................................................ 225
Data ......................................................................................................................................226
Items and Variables............................................................................................................................... 226
Violence, Victimization and Insecurity................................................................................226
Democratic Political Culture ................................................................................................226
Empirical Hypothesis ...........................................................................................................227
Victimization, Insecurity and Democratic Political Culture in the Region........................................... 228
Violence and Attitudes of Support for Democracy in Central America ............................................... 236
Conclusions........................................................................................................................................... 241
References ............................................................................................................................................. 243
XI. Corruption and its Impact on Latin American Democratic Stability ............... 251
Dominique Zphyr
Abstract ................................................................................................................................................. 251
iii
The Level of Corruption in the Region ................................................................................................. 253

Modes of Victimization by Corruption in Latin America and the Caribbean....................................... 257
Who are the most likely to become victims of corruption? .................................................................. 263
Corruption and its Impact on Democracy ............................................................................................. 266
Victimization by Corruption and Satisfaction with Democracy ..........................................267
Victimization by Corruption and Support for Democracy as an Ideal Form of Government
..............................................................................................................................................269
Victimization by Corruption and Support for the Political System .....................................271
Conclusions........................................................................................................................................... 272
References ............................................................................................................................................. 274
XII. Justice and democracy: the Rule of Law in the Americas.................................. 277
Juan Carlos Donoso
Abstract ................................................................................................................................................. 277
Judicial Power ................................................................................................................................... 277
Support for the Institution of the Rule of Law ...................................................................................... 285
Democracy and the Rule of Law........................................................................................................... 293
Conclusion ............................................................................................................................................ 297
References ............................................................................................................................................. 298
iv
Preface
The United States Agency for International Development (USAID) takes pride in its
support of the Latin American Public Opinion Project (LAPOP) democracy and governance
surveys in Latin America and the Caribbean over the past two decades. LAPOP findings have
been a crucial tool to USAID missions in diagnosing the nature of the democratic challenge;
sparking policy dialogue and debate within Latin American countries; monitoring on-going
USAID programs; and evaluating and measuring USAID performance in supporting democracy
and good governance in the region. The reports have often served as the voice of citizens on
the quality of democracy. We hope that this 2006-2007 study also proves to be useful to policy-
makers, democracy advocates, donors and practitioners.
The decision to undertake democracy surveys in Latin America and the Caribbean
emerged from the USAID country missions, where field democracy officers have increasingly
depended on them as a management and policy tool. The depth and breadth of the questionnaire
allows us to look beyond simple questions and examine complex relationships related to gender,
ethnicity, geography, economic well-being, and other conditions, and delve deeply into specific
practices and cultures to identify where our assistance might be most fruitful in promoting
democracy. The surveys represent a unique USAID resource, as a comparative, consistent, and
high quality source of information over time. USAID is grateful for the leadership of Dr.
Mitchell Seligson at Vanderbilt University, his outstanding Latin American graduate students
from throughout the hemisphere and the participation and expertise of the many regional
academic and expert institutions that have been involved in this project.
Two recent trends in these surveys have made them even more useful. One is the addition
of more countries to the survey base, using a core of common questions, which allows valid
comparisons across systems and over time. The second, and even more important, is the
introduction of geographically or project-based over-sampling in some of the countries where
USAID has democracy programs. The result is a new capability for USAID missions to examine
the impact of their programs in statistically valid ways by comparing the before and after of our
work, and also comparing changes in the areas where we have programs to changes in areas
where we do not have them. These methodologies should provide one of the most rigorous tests
of program effectiveness of donor interventions in any field.
Promoting democracy and good governance is a US government foreign policy priority,

and our investment of both effort and money is a substantial one. Democratic development is a
relatively new field of development, however, and our knowledge of basic political relationships
and the impact of donor assistance is still at an early phase. It is critical that we be able to
determine which programs work and under what circumstances they work best, learning from our
experience and constantly improving our programs. To meet this challenge, USAID has
undertaken a new initiative, the Strategic and Operational Research Agenda, (SORA). With the
assistance of the National Academy of Sciences. SORA has already incorporated the insights of
numerous experts in political science and research methodology into our work. The LAPOP
democracy surveys are a critical component of this evaluation effort. We hope their findings will
v
stimulate a dialogue among governments, NGOs, scholars and the public that will help, in the
long run, to solidify democracy in Latin America.
Dr. Margaret Sarles

Division Chief, Strategic Planning and Research
Office of Democracy and Governance
U.S. Agency for International Development
vi
Foreword
The AmericasBarometer, 2006-2007: Background to the Study
By Mitchell A. Seligson
Centennial Professor of Political Science
and Director, the Latin American Public Opinion Project (LAPOP)
Vanderbilt University
I am very pleased to introduce to you the 2006-2007 round of the AmericasBarometer

series of surveys, one of the many and growing activities of the Latin American Public Opinion
Project (LAPOP). That project, initiated over two decades ago, is hosted by Vanderbilt
University. LAPOP began with the study of democratic values in one country, Costa Rica, at a
time when much of the rest of Latin America was caught in the grip of repressive regimes that
widely prohibited studies of public opinion (and systematically violated human rights and civil
liberties). Today, fortunately, such studies can be carried out openly and freely in virtually all
countries in the region. The AmericasBarometer is an effort by LAPOP to measure democratic
values and behaviors in the Americas using national probability samples of voting-age adults.
The first effort was in 2004, when eleven countries were included, and all of those studies are
already available on the LAPOP web site. The present study reflects LAPOPs most extensive
effort to date, incorporating 20 countries. For the first time, through the generosity of a grant
from the Center for the Americas, it was possible to include the United States and Canada. The
United States Agency for International Development (USAID) provided the core funding to
enable to study to incorporate much of Latin America and the Caribbean, so that in 2006-2007, as
of this writing, the following countries have been included: Mexico, Guatemala, El Salvador,
Honduras, Nicaragua, Costa Rica, Panama, Colombia, Peru, Chile, Dominican Republic, Haiti
and Jamaica). The U.S and Canada were also included, with the support of the Center for the
Americas. Brazil and Uruguay were added too late to be included in this volume. The sample and
questionnaire designs for all studies were uniform, allowing direct comparisons among them, as
well as detailed analysis within each country. The 2006-2007 series involves a total of
publications, one for each of the countries, authored by the country teams, and a summary study,
written by the author of this Foreword, member of the LAPOP team at Vanderbilt and other
collaborators.
We embarked on the 2006-2007 AmericasBarometer in the hope that the results would
be of interest and of policy relevance to citizens, NGOs, academics, governments and the
international donor community. Our hope is that the study could not only be used to help advance
the democratization agenda, it would also serve the academic community which has been
engaged in a quest to determine which values are the ones most likely to promote stable
democracy. For that reason, we agreed on a common core of questions to include in our survey.
The United Nations Development Programme (UNDP) provided a generous grant to LAPOP t
bring together the leading scholars in the field in May, 2006, in order to help determine the best
questions to incorporate into what was becoming the UNDP Democracy Support Index. The
scholars who attended that meeting prepared papers that were presented and critiqued at the
vii
Vanderbilt workshop, and helped provide both a theoretical and empirical justification for the
decisions taken. All of those papers are available on the LAPOP web site.
The UNDP-sponsored event was then followed by a meeting of the country teams in
Heredia, Costa Rica, in May, 2006. Key democracy officers from USAID were present at the
meeting, as well as staffers from LAPOP at Vanderbilt. With the background of the 2004 series
and the UNDP workshop input, it became fairly easy for the teams to agree to common core
questionnaire. The common core allows us to examine, for each nation and across nations, such
issues as political legitimacy, political tolerance, support for stable democracy, civil society
participation and social capital, the rule of law, participation in and evaluations of local
government, crime victimization, corruption victimization, and voting behavior. Each country
study contains an analysis of these important areas of democratic values and behaviors. In some
cases we find striking similarities from country-to-country, whereas in other cases we find sharp
contrasts.
A common sample design was crucial for the success of the effort. Prior to coming to
Costa Rica, the author of this chapter prepared for each team the guidelines for the construction
of a multi-stage, stratified area probability sample with a target N of 1,500. In the Costa Rica
meeting each team met with Dr. Polibio Crdova, President of CEDATOS, Ecuador, and region-
wide expert in sample design, trained under Leslie Kish at the University of Michigan.
Refinements in the sample designs were made at that meeting and later reviewed by Dr. Crdova.
Detailed descriptions of the sample are contained in annexes in each country publication.
The Costa Rica meeting was also a time for the teams to agree on a common framework
for analysis. We did not want to impose rigidities on each team, since we recognized from the
outset that each country had its own unique circumstances, and what was very important for one
country (e.g., crime, voting abstention) might be largely irrelevant for another. But, we did want
each of the teams to be able to make direct comparisons to the results in the other countries. For
that reason, we agreed on a common method for index construction. We used the standard of an
Alpha reliability coefficient of greater than .6, with a preference for .7, as the minimum level
needed for a set of items to be called a scale. The only variation in that rule was when we were
using count variables, to construct an index (as opposed to a scale) in which we merely wanted
to know, for example, how many times an individual participated in a certain form of activity. In
fact, most of our reliabilities were well above .7, many reaching above .8. We also encouraged all
teams to use factor analysis to establish the dimensionality of their scales. Another common rule,
applied to all of the data sets, was in the treatment of missing data. In order to maximize sample
N without unreasonably distorting the response patterns, we substituted the mean score of the
individual respondents choice for any scale or index in which there were missing data, but only
when the missing data comprised less than half of all the responses for that individual.
Another agreement we struck in Costa Rica was that each major section of the studies
would be made accessible to the layman reader, meaning that there would be heavy use of
bivariate and tri-variate graphs. But we also agreed that those graphs would always follow a
multivariate analysis (either OLS or logistic regression), so that the technically informed reader
could be assured that the individual variables in the graphs were indeed significant predictors of
the dependent variable being studied. We also agreed on a common graphical format (using chart
viii
templates prepared by LAPOP for SPSS 14). Finally, a common informed consent form was
prepared, and approval for research on human subjects was granted by the Vanderbilt University
Institutional Review Board (IRB). All senior investigators in the project studied the human
subjects protection materials utilized by Vanderbilt and took and passed the certifying test. All
publicly available data for this project are deeidentified, thus protecting the right of anonymity
guaranteed to each respondent. The informed consent form appears in the questionnaire appendix
of each study.
A concern from the outset was minimization of error and maximization of the quality of
the database. We did this in several ways. First, we agreed on a common coding scheme for all
of the closed-ended questions. Second, our partners at the Universidad de Costa Rica prepared a
common set of data entry formats, including careful range checks, using the U.S. Census
Bureaus CSPro software. Third, all data files were entered in their respective countries, and
verified, after which the files were sent to LAPOP at Vanderbilt review. At that point, a random
list of 100 questionnaire identification numbers was sent back to each team, who were then asked
to ship those 100 surveys via express courier LAPOP for auditing. This audit consisted of two
steps, the first involved comparing the responses written on the questionnaire during the
interview with the responses as entered by the coding teams. The second step involved comparing
the coded responses to the data base itself. If a significant number of errors was encountered
through this process, the entire data base had to be reentered and the process of auditing was
repeated on the new data base. Fortunately, in very few cases did that happen in the 2006-2007
AmericasBarometer. Finally, the data sets were merged by our expert, Dominique Zphyr into
one uniform multi-nation file, and copies were sent to all teams so that they could carry out
comparative analysis on the entire file.
An additional technological innovation in the 2006-2007 round is that we used handheld

computers (Personal Digital Assistants, or PDAs) to collect the data in five of the countries. Our
partners at the Universidad de Costa Rica developed the program, EQCollector and formatted it
for use in the 2006-2007 survey. We found this method of recording the survey responses
extremely efficient, resulting in higher quality data with fewer errors than with the paper-and-
pencil method. In addition, the cost and time of data entry was eliminated entirely. Our plan is to
expand the use of PDAs in future rounds of LAPOP surveys.
The fieldwork for the surveys was carried out only after the questionnaire were pretested
extensively in each country. In many cases we were able to send LAPOP staffers to the countries
that were new to the AmericasBarometer to assist in the pretests. Suggestions from each country
were then transmitted to LAPOP at Vanderbilt and revisions were made. In most countries this
meant now fewer than 20 version revisions. The common standard was to finalize the
questionnaire on version 23. The result was a highly polished instrument, with common
questions but with appropriate customization of vocabulary for country-specific needs. In the
case of countries with significant indigenous-speaking population, the questionnaires were
translated into those languages (e.g., Quechua and Aymara in Bolivia). We also developed
versions in English for the English-speaking Caribbean and for Atlantic coastal America, as well
as a French Creole version for use in Haiti and a Portuguese version for Brazil. In the end, we had
versions in ten different languages. All of those questionnaires form part of the
ix
www.lapopsurveys.org web site and can be consulted there or in the appendixes for each country
study.
Country teams then proceeded to analyze their data sets and write their studies. When the
drafts were ready, the next step in our effort to maximize quality of the overall project was for the
teams to meet again in plenary session, this time in Santo Domingo de Santo Domingo, Costa
Rica. In preparation for that meeting, held in November June 2004, teams of researchers were
assigned to present themes emerging from the studies. For example, one team made a
presentation on corruption and democracy, whereas another discussed the rule of law. These
presentations, delivered in PowerPoint, were then critiqued by a small team of our most highly
qualified methodologists, and then the entire group of researchers and USAID democracy staffers
discussed the results. That process was repeated over a two-day period. It was an exciting time,
seeing our findings up there in black and white, but it was also a time for us to learn more about
the close ties between data, theory and method. After the Costa Rica meeting ended, the draft
studies were read by the LAPOP team at Vanderbilt and returned to the authors for corrections.
Revised studies were then submitted and they were each read and edited by Mitchell Seligson, the
scientific coordinator of the project., read and critiqued each draft study. Those studies were then
returned to the country teams for final correction and editing, and were sent to USAID democracy
officers for their critiques. What you have before you, then, is the product of the intensive labor
of scores of highly motivated researchers, sample design experts, field supervisors, interviewers,
data entry clerks, and, of course, the over 27,000 respondents to our survey. Our efforts will not
have been in vain if the results presented here are utilized by policy makers, citizens and
academics alike to help strengthen democracy in Latin America.
Acknowledgements
The study was made possible by the generous support of the United States Agency for
International Development (USAID). Margaret Sarles, Chief of Strategic Planning and Research
in the Office of Democracy and Governance of USAID, assisted by Eric Kite, Maria Barrn and
Elizabeth Ramirez in the Latin American Bureau, secured the funding and made possible the
entire project thanks to their unceasing support. All of the participants in the study are grateful to
them. At Vanderbilt University, the study would not have been possible without the generosity,
collaboration and hard work of many individuals. Vanderbilts Dean of Arts and Science, Richard
MacCarty provided financial support for many critical aspects of the research. Nicholas S.
Zeppos, Provost and Vice Chancellor for Academic Affairs generously offered LAPOP a suite of
offices and conference space, and had it entirely reconditioned and equipped for the project. Vera
Kutzinski, Director of the Center for the Americas has strongly supported the project
administratively and financially, and contributed key funding to enable the inclusion of the
United States and Canada in this round of the AmericasBarometer. Her administrative assistant,
Janelle Lees made lots of things happen efficiently. Neal Tate, Chair of the Department of
Political Science at Vanderbilt has been a strong supporter of the project since its inception at
Vanderbilt and facilitated its integration with the busy schedule of the Department. Tonya Mills,
Grants Administrator and Patrick D. Green, Associate Director, Division of Sponsored Research,
Vanderbilt University performed heroically in managing the countless contract and financial
details of the project. In a study as complex as this, literally dozens of contracts had to be signed
and hundreds of invoices paid. They deserve my special appreciation for their efforts.
x
At LAPOP Central, the burden of the project fell on Pierre Martin Dominique Zphyr, our
LAPOP Research Coordinator and Data Analyst. Dominique worked tirelessly, almost always
seven days a week, on virtually every aspect of the studies, from their design through their
implementation and analysis. He also had central responsibility for preparing the training
material for the teams for the data analysis and for handling the data audits and merging of the
data bases. Dominique also served as Regional coordinator of the Caribbean countries, and
personally did the pretesting and interviewer training in each of them. Finally, he worked as co-
collaborator on the Haiti study. Julio Carrin of the University of Delaware served as Regional
Coordinator for Mexico, Central America and the Andes. He managed this while also serving as
co-collaborator of the Peru study. The members of the LAPOP graduate research team were
involved in every aspect of the studies, from questionnaire design, data audits and overall quality
control. I would like to thank them all: Mara Fernanda Boidi, Abby Crdova Guilln, Jos
Miguel Cruz, Juan Carlos Donoso, Jorge Daniel Montalvo, Daniel Moreno Morales, Diana Orces,
and Vivian Schwarz-Blum. Their Ph.D. programs at Vanderbilt are being supported by USAID,
the Vanderbilt University Center for Latin American and Iberian Studies and the Department of
Political Science. My colleague Jon Hiskey participated in our weekly meetings on the surveys,
adding his own important expertise and encouragement. Our web master, Mara Clara Bertini,
made sure that our efforts were transparent, and has done an outstanding job managing the ever-
growing web page of LAPOP and the AmericasBarometer. Hctor Lard and Roberto Ortiz were
responsible for cover design and text formatting, and did so with great attention to detail.
Critical to the projects success was the cooperation of the many individuals and
institutions in the countries studied who worked tirelessly to meet what at times seemed
impossible deadlines. Their names, countries and affiliations are listed below:
Country Researchers
Summary Prof. Mitchell Seligson, Director of LAPOP, and Centennial Professor of Political Science,
Report Vanderbilt University (project director)
Mexico and Central America Group
Mexico Dr. Kenneth M. Coleman, Senior Research Analyst and Study Director of Market Strategies, Inc.
Pablo Pars Garca, President of DATA Opinin Pblica y Mercados
Guatemala Dr. Dinorah Azpuru, Senior Associate at ASIES in Guatemala and Assistant Professor of
Political Science at Wichita State University in the U.S.
El Salvador Dr. Ricardo Crdova (Salvadoran national), President of FundaUngo, El Salvador
Prof. Miguel Cruz, Director of IUDOP (Public Opinion Institute) at the Universidad
Centroamericana (UCA)
Honduras Prof. Miguel Cruz, Director of IUDOP (Public Opinion Institute) at the Universidad
Centroamericana (UCA)
Jos Rene Argueta, Ph.D. candidate, University of Pittsburgh
Nicaragua Prof. Manuel Ortega-Hegg, Director of the Centro de Anlisis Socio-Cultural (CASC) at the
Universidad Centroamericana (UCA), Managua, Nicaragua
Marcelina Castillo Venerio, Centro de Anlisis Socio-cultural (CASC), Universidad
Centroamericana. (UCA)
Costa Rica Dr. Luis Rosero, Director of Centro Centroamericano de Poblacin (CCP, and Professor at the
Universidad de Costa Rica.
Jorge Vargas, Sub-Director of the Estado de la Nacin project, United Nations
Panama Dr. Orlando Prez, Associate Professor of Political Science at Central Michigan University
xi
Country Researchers
Andean/Southern Cone Group
Colombia Prof. Juan Carlos Rodrguez-Raga, Professor at the Universidad de los Andes
Ecuador Dr. Mitchell Seligson, Director of LAPOP, and Centennial Professor of Political Science,
Juan Carlos Donoso, Ph.D. student, Vanderbilt University
Daniel Moreno, Ph.D. student, Vanderbilt Universtity
Diana Orcs, Ph.D. student, Vanderbilt University
Vivian Schwarz-Blum, Ph.D student, Vanderbilt University
Peru Dr. Julio Carrin (Peruvian national) Associate Professor at the University of Delaware in the
US, and Researcher at the Instituto de Estudios Peruanos
Patricia Zrate Ardela, Researcher at the Instituto de Estudios Peruanos
Bolivia Dr. Mitchell Seligson Director of LAPOP, and Centennial Professor of Political Science,
Abby B. Crdova, Ph.D. student, Vanderbilt University
Juan Carlos Donoso, Ph.D. student, Vanderbilt University
Daniel Moreno, Ph.D. student, Vanderbilt Universtity
Diana Orcs, Ph.D. student, Vanderbilt University
Vivian Schwarz-Blum, Ph.D. student, Vanderbilt University
Paraguay Manuel Orrego, CIRD, Paraguay
Chile Juan Pablo Luna, Instituto de Ciencia Poltica, Pontificia Universidad Catlica de Chile
Uruguay Mara Fernanda Boidi, Ph.D. student, Vanderbilt University
Dr. Mara del Rosario Queirolo, Professor of Political Science at the Universidad de Montevideo
Brazil Denise Pavia, Universidade Federal de Gois, Gois, Brazil
Simon Bohn, York University
Rachael Meneguello, Brazil, Diretora do Centro de Estudos de Opinio Pblica (CESOP) and
Professor of Political Science, University of Campinas, Brazil
David Samules, University of Minnesota
Luicio Renno, University of Arizona
Caribbean Group
Dominican Dr. Jana Morgan Kelly Assistant Professor of Political Science at the University of Tennessee in
Republic the US
Dr. Rosario Espinal, Professor of Sociology Science at Temple University in the US
Guyana Dr. Mark Bynoe, Director, School of Earth and Environmental Sciences, University of Guyana
Ms. Talia Choy, Lecturer, Department of Government and International Affairs, University of
Guyana.
Haiti Dominique Zephyr, Research Coordinator of LAPOP, Vanderbilt University
Yves Franois Pierre, Groupe de Recherche en Sciences Sociales (GRESS)
Jamaica Ian Boxill, Professor of Comparative Sociology, Department of Sociology, Psychology and
Social Work, UWI, Mona.
Roy Russell, Lecturer in statistics, Department of Sociology, Psychology and Social Work, UWI,
Mona.
Arlene Bailey, Information Sytems specialist, Department of Sociology, Psychology and Social
Work, UWI, Mona.
Balford Lewis, Lecturer in research methods, Department of Sociology, Psychology and Social
Work, UWI, Mona.
LLoyd Waller, Lecturer in research methods, Department of Government, UWI, Mona
Finally, we wish to thank the more than 30,000 individuals in these countries who took
time away from their busy lives to answer our questions. Without their cooperation, this study
would have been impossible.
Nashville, Tennessee, March, 2008
xii
Overview
By Margaret Sarles and Mitchell A. Seligson
There is much to celebrate in the strengthening of democratic practices and institutions in

Latin America. Once ruled almost entirely by dictators, countries in the region are now ruled
almost entirely by democrats. Elections have emerged as the established norm for deciding who
rules, replacing the frequent and irregular transfers of power via military and executive coups of
the past. Indeed, even the most charismatic leaders who espouse radical change seem committed
to free and fair elections as the path to legitimate leadership. With greater access to power
through elections, even the radical left is using the political system to come to power, rather than
using violence. The holdouts, such as the protracted guerrilla movement in Colombia, or the
nearly half-century long socialist dictatorship in Cuba, seem like anachronistic, isolated
anomalies in a region committed to electoral politics. Similarly, democratic trends are evident
outside of electoral politics. Civil society organizations and independent business groups grow
and prosper; political parties take their turns in and out of power; an increasingly independent and
feisty press uncovers malfeasance, and local officials assert more independence from the center.
State-sanctioned human rights violations, so frequent in the recent past, are now the exception to
the rule. On the governance side, Latin American democracies have been able to put together
political coalitions to successfully tackle a number of important policy issues, providing
macroeconomic stability, increased investment, increased access to education and health, and a
number have made impressive gains in fighting poverty. Democracy seems to have become, to
use Adam Przeworskis classic terminology, the only game in town.
Yet, continued progress in the direction of democratic consolidation and higher quality
democracies is certainly not a foregone conclusion, as the results of Freedom Houses 2007
Freedom in World survey demonstrates. The report notes an overall decline in the state of
freedom worldwide, and a growing pushback against democracy by authoritarian governments,
including Venezuela. Even in the relatively consolidated democracies of Mexico, Argentina, and
Brazil, freedom scores declined last year. While competitive electoral politics continue to
thrive, and the institutionalization of electoral institutions is indeed a significant hallmark of
progress, other democratic institutions show worrisome signs of weakness in the region.
Continuing government corruption scandals illustrate the weakness of accountability and rule of
law, as do high crime rates. Furthermore, after being democratically elected, a number of
presidents are pushing the boundaries of executive power to new extremes, making every effort to
amass greater and greater power at the expense of representative institutions fundamental to
liberal democracy. Political parties in many countries remain weak. The rule of law is fragile in
many countries, while local governments are starved for resources. Ethnic-based parties are on
the rise in some countries, suggesting increasing sectarianism.
What are we to make of this mixture of trends, both positive and negative? At the macro-
level, the patterns are not clear. At best, they show that democratic consolidation is not a linear
process, providing evidence of democratization as a somewhat lumpy process, with some aspects
proceeding well, while others seem frustratingly intractable. It is not clear whether a negative
change in a particular year, in a particular country, is a real set-back to democracy, part of an
xiii
underlying movement away from liberal democracy, or whether we are simply seeing an
unsteady, back and forth process that may in fact be compatible with long-term democratic
consolidation. To understand Latin American democracies, we need to go below the macro level.
We need more and better data. And we need to undertake a more fine-grained analysis that
permits detailed comparisons among specific aspects of democratic change in the region.
It is the purpose of this regional volume to begin such a comparative, detailed analysis of
democratic processes in Latin America and the Caribbean, to provide a different kind of lens
through which to analyze and make judgments on the state of democracy in the region.
This research is based on the AmericasBarometer democracy surveys undertaken in 22

countries in 2006 and 2007, involving more than 34,000 in-depth, face-to-face interviews with
citizens in the region. The AmericasBarometer data reveal what citizens do, think, and feel in
terms of democracy, as well as information on personal attributes, allowing detailed analysis
available through no other source. For nearly all of the countries, detailed country case studies
have been written based on the surveys, covering themes such as democratic legitimacy,
corruption, crime and violence, and citizen participation.1 For some countries, there are multiple
surveys, stretching back ten years or more. As they are published, the country survey studies
often receive great publicity, and their results are debated and discussed in the national press and
among people interested in politics, contributing to the overall dialogue on the state of democracy
and democratic reforms.
This regional report is a companion to the country studies. It brings together some of the
most important findings from the surveys, but focuses on the region as a whole. The articles
cover topics critical to democratic development and sustainability in the region, and should
broadly contribute to the ongoing discourse on democracy and governance reform in policy and
development communities.
We have chosen to focus on a single, important theme: the challenges to democracy in

Latin America and the Caribbean. Even focused this way, we have nonetheless had to be
selective in choosing a limited number of topics. We selected them on the basis of 1) our
evaluation of their importance to the processes of democratic development; 2) their salience
within Latin America as important political topics; and 3) how well survey research can deal with
the topic. Obviously, there are many aspects of institutional development, macro-political
change, and leadership that require analysis beyond survey research that are important for
understanding democratic development in Latin America. And, conversely, there are certain
themes for which surveys provide excellent, and sometimes the only, robust empirical evidence.
We have naturally emphasized those themes in this report.
1
The country reports are available on the Latin American Public Opinion Project website in pdf format. (see
www.LapopSurveys.org). The data sets for each country are also available world-wide, free, for on-line analysis
through the above website. We welcome and encourage other scholars to use the data freely, both to test the findings
presented here and to undertake their own research. Given the sheer quantity of data available, this volume, as well
as the country monographs, can only begin to scratch the surface in uncovering interesting findings on democracy.
xiv
The articles in this volume cover a broad spectrum of approaches to democracy, but
present a unified set of analyses in many ways. All link democratic theory to some aspect of the
survey data, and all suggest the broad policy concern that lies behind the topic. Most deal with
some aspect of the questions of who supports democracy, what kind of democracy, and under
what conditions do they support it. Some search for broad findings, while others dig more deeply
into a particular relationship that has implications for democratic sustainability. Many of them
combine survey data with macro, structural variables, using complex regression analysis that
allow us to assess the relative importance of individual attitudes and behaviors compared to
country-level characteristics such as a countrys wealth. As a group, they show how much can be
learned from surveys that goes beyond easy characterizations of the regions democratic culture.
In fact, many of them present findings that directly contradict what has passed for the common
wisdom about Latin American democracies.
Careful readers will note that there is not total agreement on all points among the authors.
In part, this reflects the fact that authors have often chosen different, but related, concepts and
measurements to examine a basic relationship, and the findings themselves therefore will differ.
In part, however, it reflects the imperfect state of democratic theory itself, in which there is still
great debate over the significance of unresponsive political parties, or peoples willingness to
consider political options other than democracy as a political system. We have not made an
effort to have all the authors conform to a single definition of democratic stability or change, and
in fact welcome the diverse interpretations that they bring to their work.
These articles as a whole move the debate on democratization away from the tendency of
policy wonks to declare the regions progress in black or white terms, or to provide facile
characterizations of an entire hemisphere. While the aggregation of country level data to a
regional level surely leads to some loss of nuance, the authors are careful to limit their
generalization, or provide country level comparisons, when appropriate. Overall, these region-
based analyses provide the kind of comparative evidence that has been lacking, in the debate on
democratization.
The Challenges
The first challenge to democracy we address is one of the most worrisome to analysts
focusing on the contemporary state of democratization in Latin America: the willingness of many
Latin Americans to grant, and even encourage, their newly elected Presidents to disregard the
normal checks and balances of a modern democracy and assume semi-authoritarian powers.
Creeping presidentialism in Latin America was noted by Guillermo ODonnell in his classic
paper on delegative democracy in 1995, and the tendency seems, if anything, to have grown
stronger in the thirteen years since the article was published (O'Donnell 1995). At times it has
been the a president himself who has sought widened powers, expanding the use of decree
laws, for example, that can be enacted with limited legislative input, or none at all. Many
presidents have also successfully lobbied for Constitutional changes allowing them to stay in
power beyond their original limits. But it would be a mistake to think that presidents in these
relatively new democracies are taking on these powers without citizen approval. Such changes,
in fact, are usually very popular. It appears that many citizens want to have their free and fair
xv
elections, but that once elected, they feel comfortable with giving their presidents extraordinary
powers.
Many researchers have seen this political/cultural characteristic as an indication of the

emergence of electoral authoritarianism, or a contemporary form of democratic populism,
characterized by virtually unchecked executives, supported by mass movements who see the
opposition and competitive party politics as a threat, rather than as a necessary condition for
liberal democracy to prosper. Is this an important characteristic of democratic culture in Latin
America? If so, does it represent a limit to the institutionalization of democratic practices in the
region?
It is precisely this phenomenon that Mitchell A. Seligson explores in the first article in
this report: the extent to which citizens support executive dominance, and are willing to abandon
many of the checks and balances required in a sustainable democratic state. Seligson looks at
electoral authoritarianism in terms of the left/right political ideology of citizens. He finds those
identifying as left are more skeptical and mistrustful of political institutions, less inclined to
believe that democracy is better than other systems, and more favorable to a strong leader. If
Latin America is turning left, as a number of recent elections seem to demonstrate, this might
be a cause for great concern. It would indicate an electorate less and less committed to a
democratic political system.
Yet, Seligsons analysis of context provides a much more nuanced picture. First, he finds
that even with the small increases in leftist orientation that seem to have taken place, Latin
American citizens remain slightly to the right ideologically, compared to other regions of the
world. More importantly, however, he is able to determine important limitations that most
citizens, even those on the left, want to impose on the power of an elected leader. He notes that
even as they greatly distrust political parties, most citizens nonetheless consider them essential to
democracy: they do not want to abandon them or other political institutions. And even while
those surveyed are impatient with the slow pace of economic reform, there are indeed important
limits to the populist measures they will support to get there.
In Chapter II, Julio Carrin then examines a more theoretical, but equally important
question for those interested in the state of democracy in Latin America: What kind of
democracy do people have in mind when they say they support democracy? How deep and
consistent is their commitment? What kind of people most and least support democracy?
Carrins approach is methodologically path-breaking, particularly in the Latin American

context. His analysis is based on open-ended questions asking people to define what they mean
by democracy. These questions, developed by AfroBarometer, and administered in numerous
African countries over time, allow us for the first time to get a good sense of citizens
understanding of democracy in Latin America, and even to compare Latin American and African
countries. It should come as no surprise that mass publics do not agree on a single definition of
democracy, since even experts disagree on its meaning. Carrin applies sophisticated modeling
techniques, making it possible to control for a series of variables that have often hidden the real
relationships between support for democracy and other attributes, uncovering somewhat startling
relationships.
xvi
While Carrin finds that a majority of Latin Americans hold a view of democracy that
embraces freedom and liberty, he also demonstrates that a significant percentage refer to
government services and benefits or economic progress as part of their definition. Perhaps even
more troubling is the number of respondents who seem to have no idea what the term means.
This in itself represents a democratic challenge.
Like Seligson, Carrin finds that people who most embrace a liberal definition of
democracy have certain characteristics: they support democracy as a system to a greater extent,
and are likely to be older, to be male, to be wealthier, to live outside of rural areas, and to have
greater political knowledge. Controlling for these attributes, Carrin finds that the often-
discussed relationship between social trust and support for democracy disappears completely.
In addition, and again in contradiction to much of the prevailing analysis, he finds that whether a
person has been a victim of crime does not make him or her less likely to support democracy.
This suggests that a policy priority on lowering crime rates may not have the effect on increasing
support for democracy, at least in general terms, which some reformers had hoped for. On the
other hand, those who are most upset at government corruption and those negative about the
national economy are indeed more inclined to support non-democratic political alternatives.
In Chapter III, Mara Fernanda Boidi tackles one of the most debated and least understood
institutional aspects of democratization in the region: the state of political parties. Political
parties are essential for a modern democratic political system to function, but are often
characterized as weak in Latin America, and mistrusted by citizens. In fact, while surveys have
shown that political parties are the least trusted political institutions world-wide, Latin American
parties have the dubious honor of being least trusted even within this sorry group. What does this
mean for democracy?
Boidis contribution to this discussion lies first in the presentation of new data. She has
amassed data from the surveys by country on party identification, participation in partisan
activities, trust in political parties, and other party-related behavior voting, that has never before
been analyzed. Her chapter illustrates the richness of the survey data as it relates to political
parties, and should spur other party researchers to pursue it further. She sets out to explain why
Latin Americans have such low trust in parties, developing a multiple regression model that
includes variables based on many of the most frequently cited reasons for low trust, including
perceived government ability to combat corruption, crime, and poverty, and to reduce
unemployment. Her preliminary analyses indicate that these factors seem to have almost no
explanatory power.
Like other contributors to this volume, Boidi is struck by what appear to be

inconsistencies in citizens views of politics and democracy. She notes that not only is trust in
parties low, voters do not use parties as an information shortcut for voting choices: only one-
fifth of the respondents choose a presidential candidate based on his or her party, with wide
variations by country. She concludes that political parties are failing to provide meaningful
alternatives for leading the government; as a consequence, they do not fulfill their
representational task. Yet, she also finds that citizens reject having a democracy without parties,
and that there is a close relationship between ideology and party vote, with people who identify
xvii
more as rightist voting more for rightist parties, and the reverse, as Seligson also showed
from a somewhat different perspective. Boidis analysis offers intriguing evidence that Latin
American political parties are not merely vehicles for personal ambition for leaders, nor
patrimonial parties kept alive only through distribution of jobs and personal linkages, but are
also based in genuine policy and ideological differences.
An even more basic challenge to the sustainability of democracy than political parties is
the low level of system support than many Latin Americans express towards their new
democracies. For democracies to remain stable, citizens must express at least minimal levels of
support for their political system. Democracies that are not supported by their citizens are
vulnerable to instability and even breakdown. This problem is explored in Chapter IV by Vivian
Schwarz-Blum.
Schwarz-Blum reiterates an important initial point: In highly industrialized countries,

system support for democracy is relatively high (even if it has declined in recent years), and
basically unaffected by a countrys poor economic performance or even by a citizens own
economic situation. Unfortunately, in Latin America, she finds that this is not the case. Schwarz-
Blum uses as a definition of system support an index composed of five questions created by
LAPOP that probes whether or not the respondent is proud of his/her political system, whether
trials are fair, whether political institutions should be respected, whether basic rights are safe-
guarded, and whether the political system should be supported. This, she argues, is a more
demanding measure of support than the question of whether citizens support democracy as better
than the alternatives used in other articles in the report. By this measure, Canada and the US
stand at the highest end of the 19 countries studied, with about two-thirds of respondents
supporting their democratic systems, with Ecuador and Paraguay at the other end, with less than
40 percent system support. These findings lead to a concern for the democratic sustainability
among the countries at the low end of the scale.
In an interesting contrast to Boidis findings that economic performance does not affect
trust in parties, Schwarz-Blum finds that good economic performance does, in fact, increase
more general support for a countrys political system. Latin American citizens dissatisfied with
their countrys economic performance, or who perceive their own economic situation negatively,
support their democracies less. Understanding this linkage of democracy support to economic
progress contributes to a much wider contemporary discussion on the relationship of economic
performance to democratic sustainability, particularly during the early years of a new democracy,
and illustrates the potential for survey data to analyze it. Her results place some of the burden of
democratic success squarely on the shoulders of those who manage national economies, including
national political leaders, central banks, parties and policy makers.
Historically, political power in Latin America has been excessively centralized at the
national level, and, within the national government, in the executive branch, making authoritarian
political rule easier and more likely. Seligson addressed one major aspect of this challenge to
democracy in his discussion of executive dominance in Chapter I, revealing some continuation of
citizen support for electoral authoritarianism even a generation after new democratic systems
were constituted, although with important limitations. Daniel Montalvo addresses the other side
of over-centralization in Chapter V, carefully examining the decentralization of power from of
xviii
the national arena into municipal governments. Is decentralization one of the solutions to
developing sustainable democracies?
Montalvos work is one of the few empirical studies that apply survey data to this
important problem. Decentralization has been promoted as a permanent impediment to the re-
imposition of authoritarian rule, as the growth of independent, sub-national power centers
potentially restricts the powers of national government. It promises to place citizens closer to the
governments that influence their daily lives, and to provide wider access to and greater
opportunities for citizen participation in government, as well as government knowledge of and
sensitivity to local voters needs. Over the past generation, country after country has instituted
direct elections for mayors and municipal authorities; the final democratic government to do so
was Costa Rica, with direct mayoral elections first taking place in 2002. Yet we know little about
whether citizens in fact feel that this promise for better democracy and improved governance has
been realized.
Montalvo, in fact, finds some divergence between the grand hopes for decentralization
and its support among citizens in Latin America. Overall, peoples trust in local government is
about the same as in national governments. However, this obscures great variation within the
region, with citizens in some countries trusting much more in local government, and preferring
that services be provided locally, while in other countries just the reverse is true. The research
sets out a typology of states, based on differential levels of trust at the national and local levels. It
then probes into the historical events that have led to these differences, illustrating how survey
data can complement and strengthen more traditional forms of country-level research. Jamaicas
positive experience with local government plans, Ecuadors poor track record with stable national
governments these and other unique country experiences help explain wide country differences.
These findings suggests that there are differential opportunities for decentralization as a means of
promoting democracy across Latin America, and that programs directed at increasing
decentralization need to consider the beliefs and attitudes of citizens in each country if they are to
reach their goals. The evidence in this chapter shows that it is not a silver bullet.
In Chapter VI, Rosario Queirolo returns to two themes discussed in other papers in the
report: the relationship of economic performance to voting patterns, and whether the rise of the
left observed in voting patterns represents a threat to democracy. She tests her hypotheses using
slightly different data in the surveys, however, which both illustrates the multiple approaches
possible using the survey data and also shows that the findings of the researchers will vary to
some degree depending on the survey questions chosen. Queirolo tests competing economic
theories that have been advanced to explain why Latin Americans have elected leftist Presidents
in a number of countries. She concludes that the move to leftist government is not driven by
fatigue with the market reforms of the Washington Consensus, nor by ideology of the left, but
rather by the failure of government whatever its ideological stripe to improve economic
performance. Latin American voters have been more driven by economic outcomes than by
ideology, she concludes. Voters choose parties that have not been tainted by being in office.
As she re-focuses us on the question of whether the left poses a danger to democracy,
she brings a different set of survey instruments to bear, using scales developed by LAPOP to
measure political tolerance and support for basic liberties, in addition to the frequently used
xix
question of whether a respondent agrees with the proposition that Democracy may have its
problems, but it is better than any other form of government. Her findings diverge in provocative
ways from those of Seligson, however. While both find that the right has higher levels of support
for a democratic political system than the left, Queirolo reports that leftists score higher in terms
of political tolerance and support for civil liberties, while rightists are less inclined to support any
use of force to overthrow a government. She concludes that the left poses no danger to
democratization in Latin America: leftists are using the democratic political system rather than
rejecting it, and they support beliefs consistent with a democratic culture.
Abby Crdova Guilln turns to another challenge to democracy the often-cited lack of a
social capital in Latin America that, according to experts like Robert Putnam and Larry
Diamond, is necessary for democracy to be sustained. This is a complex topic, and Crdova
Guilln offers important insights on Latin American democratic culture in Chapter VII. Does
Latin America have the social capital needed to sustain democracy? Guilln compares Latin
American countries to the consolidated democracies of the US and Canada, looking at the two
components of social capital, interpersonal trust and civic participation. As expected, she finds
that levels of social trust in the US and Canada are higher than in Latin American countries. In
contrast, however, civic participation rates vary greatly, led by some countries, like Haiti, with
very low levels of social trust. On the whole, the evidence shows that it is trust, rather than
participation, with the more important relationship to democratic development and social capital.
Crdova Guilln then examines the sources of social trust to determine how enduring or,
conversely, how malleable, social trust is as a societal value. If, as some have argued, it is deeply
rooted in the history and religion of the region, and resistant to change, then we can expect that
many Latin American countries will have great difficulty in sustaining democracy. Conversely, if
the concept is rooted in structural or individual factors amenable to change, then it is more likely
that if these factors are changed, a more democratic culture could take root.
She uses a hierarchical linear regression model to analyze individual and country
contextual factors together, controlling for age, gender, schooling levels, and crime victimization,
all possible explanations of differences between Latin America and the US and Canada scores on
social trust. She concludes finally that it is not individual differences, but economic country-level
factors that best explain trust levels. Like other researchers, she finds that the wealthier a
country, the higher the level of social trust. However, she finds this potentially positive
relationship between development and social trust undercut by another factor: economic
inequality. The greater the inequality, the less impact that economic growth has on social trust.
In other words, for social trust to increase, it is not enough to focus on economic development:
there must also be a focus on improving economic equality. This conclusion on the primacy of
economics to explain social trust is also important because economic development, and even,
income distribution, are malleable; they are not individual traits resistant to change. Social
trust can be changed.
To what degree does ethnicity pose a challenge to the consolidation of democracy in the
region? Daniel Moreno addresses this challenge in Chapter VIII. In many Latin American
countries, indigenous people and minorities have historically faced brutal discrimination, denied
access to political and social power. The new democracies have been more inclusive, offering
xx
political leadership opportunities for the first time to many such groups. But given Latin
American history, do traditionally excluded groups truly feel part of the new political system?
Are they willing to play the democratic political game? Will countries divide on the basis of
ethnicity?
Moreno introduces the concept of a sense of national integration, a combination of pride

in country, and a belief that ones country holds a set of common values. He argues persuasively
that this positive sense of country and political community is essential for consolidating
democracy. He then asks the question whether people from traditionally excluded groups have
the same sense of nationhood, the same sense of pride and common values, as the rest of the
population. Given historical patterns of political and social exclusion of minorities, do they
represent an unintegrated, more alienated group within the polity?
His analysis finds the answer is no, with a very few exceptions. He shows that a very
high percentage of all citizens across the hemisphere feel pride in their countries and that
minorities express the same level of pride as their fellow citizens, except in four countries,
Canada, Honduras, Peru and Panama. When he turns to the question of sharing common
values, however, some differences appear. People who describe themselves as ethnic minorities
are less persuaded that there are commonly shared values, a concern because those who believe
there are common values among citizens are also more politically tolerant. Like other researchers
in this report, he notes how important individual attributes are for explaining differences, with
older people feeling more national pride than younger ones, and women having a stronger sense
of national common values than men. Overall, the story is a positive one. Ethnic identity seems
to have no effect on the strength of the bond between citizens and the political community.
This research, as Moreno himself notes, only scratches the surface of what can be learned
about ethnicity and democracy from the survey data. Studies of voting patterns and political
behavior, in addition to political and practices, should in the future give us a better understanding
of whether the new democracies will be able to integrate all citizens fully into politics and
society.
Democratic political systems provide a means for all citizens to participate in politics,
protecting even unpopular groups from the potential tyranny of the majority, with free speech and
access to political power. Do the political cultures of Latin American countries support this very
basic principle of democracy? Diana Orcss analysis of this question in Chapter IX should lead
to serious concern among democratic reformers.
Orcs uses a very basic definition of political tolerance developed by Seligson that asks
respondents if they support the right of groups they dont agree with to vote, to demonstrate
peacefully, to run for office, and to make political speeches. Each of these is a protected right in
a democracy. Yet, many, and often most, citizens in Latin American countries do not agree that
opposition groups should have these rights. Even in the consolidated democracies of the US and
Canada, fully 20 percent of respondents are willing to deny opponents these rights. Most
countries in Latin America and the Caribbean have about 40 percent who do not agree, and in
four countries Bolivia, Honduras, Ecuador, and Panama a majority of citizens do not agree.
xxi
There are important differences in levels of tolerance among countries, with the data
providing some support for the argument that the British and French colonial experiences might
have provided a firmer base than Spanish colonialism for developing tolerance. But as she
explores the roots of political tolerance, Orcs finds one factor above all others that explain
differences in tolerance: the individuals level of education. Those with a secondary education
are much more likely to be politically tolerant, and those with a higher education even more so.
She finds other factors also at play: Women on the whole are less politically tolerant than men,
urbanites are more tolerant than those who live in rural areas. But higher education is by far the
strongest attribute of those who support basic rights for groups they dislike. These differences
are important to understand from a policy perspective, as they help guide efforts to develop a
more democratic culture.
Does crime undermine democratic sustainability? The LAPOP survey analyses have led
the way in identifying high crime victimization rates as a major challenge to democracy in Latin
America. Particularly in Guatemala, El Salvador and Honduras, citizens in repeated surveys have
said they would support a military coup dtat more to bring down crime. Some Central
American countries have among the highest crime rates in the world. As crime has soared and
youth gangs expanded, democratic reformers have struggled with how high a priority to place on
lowering crime rates as a means of improving the quality and sustainability of democracy.
Jos Miguel Cruzs findings in Chapter X may change the way reformers look at the
relationship of crime to democracy. Cruz looks at three different crime-related independent
variables: whether an individual has been a crime victim; whether he/she feels insecure due to
crime; and how he/she assesses government performance in combating crime. (It should be noted
that this is the only article that focuses on a specific government performance measure, although
the surveys include questions on government performance in combating poverty and corruption,
and is interesting for that reason alone). He also looks at a broader array of democracy indicators
as dependent variables: not only system support, but rejection of authoritarianism, and whether
the police should be allowed to ignore the law in combating crime.
The results are startling. Based on a comparison of Central American countries, Cruz
finds that crime victimization may not have the importance we have often given it, as a risk
factor for democracy. Assessment of the governments performance, not actual crime
victimization, and not even perception of insecurity, turns out to be the most significant crime-
related independent variable: he notes that personal victimization loses its powers of prediction
in countries where it had previously appeared to be of importance, in Honduras and Nicaragua.
It is institutional performance that is the key.
There is a complex relationship between victimization, perceptions of insecurity, and

perceptions of institutional performance that needs greater exploration. His work also brings out
some intriguing relationships between assessment of performance and peoples willingness to let
the police violate the law, and other non-democratic attitudes. Further research on this topic is
likely to improve our ability to develop reforms to fight crime in a better way, which strengthens
the foundations of a democratic culture rather than potentially undermining it.
xxii
In Chapter XI, Dominique Zephyr addresses the challenge that governmental corruption
poses to democratic stability in the region. Democracy analysts often refer to crime and
corruption as a double-barreled shotgun that weakens the new democracies and turns citizens
away from them. As Zephyr quotes a former president of the Organization of American Studies:
Corruption is a terrible cancer that threatens the legitimacy of institutions and the rule of law.
Zephyr uses the empirical data on corruption developed by Mitch Seligson with the
LAPOP surveys over the past decade, and expands the analysis to include many more countries.
In general, analysts have needed to rely on Transparency Internationals Corruption Perception
Index as the best measure available of a countrys corruption levels, and it is still the only world-
wide measure available. Within Latin America, however, the AmericasBarometer now makes
available a series of questions to determine whether a citizen has had to pay a bribe, and to
whom. This corruption victimization index moves us away from perceptions and contagion
effects, towards a much harder and more reliable measure of corruption. Using the LAPOP data,
then, do the high corruption levels in many Latin American countries undermine democracy?
Zephyr notes that previous studies using TI data found no evidence that corruption
affected citizens opinions of democracy as a form of government. Using the LAPOP corruption
victimization data, however, a different picture emerges. In fact, citizens who are forced to pay a
bribe have less support for democracy as a system, trust government less, and view their
democratic governments as less legitimate than people who have not have to pay a bribe.
Wealthier, older people are more likely to have to pay bribes people who in general are greater
supporters of their democratic governments and of democracy in general. Corruption levels vary
greatly by country, and within country, by the source of corruption police systems, health
ministries, local government, etc. As with many of the topics in this report, Zephyrs work raises
a whole series of subsidiary questions for future research that would be useful not only for
analysts but for practical policy-makers who need to focus their resources on where the problem
resides.
In the final chapter, Juan Carlos Donoso looks at the challenge of rule of law in the
region. Rule of law means many things to many people, and no single set of empirical data can
capture its multi-dimensionality. But Donoso rightly identifies the need to shed whatever light is
possible on a part of democracy that all bemoan as weak, in which donors have put great amounts
of funding, and which suffers from lack of valid and reliable measurement.
This chapter brings together some very interesting and cutting-edge measures of rule of
law for the first time. To capture whether the judiciary is independent, Donoso adopts a measure
developed from the State Departments Annual Human Rights report. This State Department
report has assumed rather remarkable credibility over the past decade or more, and political
scientists are increasingly constructing measurable variables from its annual country essays on
rule of law and human rights. He analyzes the autonomy of the judiciary in terms of peoples
support for democracy, their level of social trust, and adds some specific questions on whether
citizens believe that trials are fair, that they can trust the police and trust the Supreme Court.
He arrives at some generally discouraging conclusions. The Supreme Court is one of the
least trusted institutions in Latin America, although there are great differences among countries.
xxiii
In fact, even controlling for alternative explanations, there is an independent country effect in
rule of law. In Ecuador, where the Supreme Court has been the football in a no-holds-barred
political struggle for years, citizens have lower trust in the Court than similar citizens elsewhere
in the region. Most individual characteristics are not important, except for education and whether
a person has been victimized by corruption or crime. These have a corrosive effect on trust in
rule of law institutions.
Donoso makes a compelling case that attitudes towards rule of law are crucial for
sustaining a democracy. In countries where the rule of law is more entrenched, citizens seem to
recognize this, have greater trust in institutions, and more support for democracy. When citizens
have less trust in rule of law institutions, they tend to agree with non-democratic statements such
as Judges get in the way of our presidents and they should be ignored.
Conclusions
These chapters provide much food for thought in how to understand democratic culture,
and its importance for sustaining the new democracies of Latin America. They cover topics often
not addressed in the literature, and make use of rich data that increase our knowledge of
democracy below the surface of current events, crises, and elections. This is a detailed, nuanced
micro-level look at democracy in the Americas. The findings present a complex mosaic,
suggesting that democracy itself is a complex phenomenon, all too often simplified by a single
index or number. The data behind these chapters in fact pose a challenge to democratic theorists:
our information may be out-running our political constructs. Furthermore, policy makers and
developmentalists will find in the information provided here conclusions that may challenge the
conventional ways reformers have focused on democratic development. We look forward to
widening the discussion on Latin Americas path of democratic reform.
References
Huntington, Samuel P. Political Order in Changing Societies. New Haven: Yale University
Press, 1968.
O'Donnell, Guillermo. "Delegative Democracy." Journal of Democracy 5, no. 1 (1995): 55-69.
xxiv
Part I. Challenges to
the Institutional
Infrastructure of
Democracy
I. The Rise of Populism and the Left: Challenge to Democratic

Consolidation?1
Mitchell A. Seligson
Abstract
What are we to make of the recent dramatic rise of the left in power and the resurrection of
populism in Latin America? Do the trends reflect little more than a pendular shift in mood
among voters, or are there more profound implications for democratic development? This paper
examines citizens political orientation, showing that over the last two years, there has been a
slight movement to the left, which is reflected at the ballot box. Furthermore, those with a
more leftist orientation are more skeptical and less supportive of traditional political institutions
and democratic values, and more willing to support a non-elected political leader. While levels of
political tolerance, a critical democratic value, are about the same across the ideological
spectrum, on most measures, those with on the right are firmer in their support for checks and
balances on executive power and less willing to grant unlimited powers to a President. The
variance across the countries is the region is high, however, and it is not clear whether these
relationships will endure. However, the willingness of citizens to grant their Presidents
extraordinary power should remain a concern to those committed to democratic consolidation.
A growing number of countries in Latin America now have elected Presidents from leftist
or center-left political parties, more than at any time in history. In South America, leftist-oriented
Presidents now lead most countries, including every large one: Argentina, Bolivia, Brazil, Chile,
Ecuador, Guyana, Peru, Uruguay, and Venezuela, with the possibility of a leftist victory in
Paraguay next year. While most of the Presidents of these countries publically express a
commitment to democratic values, some are pressing the limits of executive dominance
acceptable in a democracy, and may have gone beyond it. Venezuelas Hugo Chavez, who
attempted a military coup in 1992, and, has now been popularly elected in two elections, has
openly challenged the principles of liberal democracy, controlling the courts and the legislature,
openly espousing single-party rule, and embracing Castro Cuba as a political model. Two more
recent additions to the left, Bolivias Evo Morales, who took office in 2006, and Ecuadors Rafael
Correa who began his presidency in early 2007, are, in various ways, also challenging the checks-
and-balances system of liberal democracy, supporting Constitutional reforms to enhance their
powers.
1
An earlier draft of this paper was presented at Brigham Young University, March 15, 2007, and a shorter,
somewhat different version was published in the Journal of Democracy, The Rise of the Left and Populism in Latin
America, volume 18, No. 3 (July, 2007), pp. 81-95. I would like to thank the Kirk Hawkins and Darren Hawkins of
BYU for their comments, as well as the suggestions made by many at LAPOP, especially Jos Miguel Cruz and John
Booth.
Centennial Professor of Political Science and Director, Latin American Public Opinion Project, Vanderbilt
University.
3
Of course, it is not only leftist Presidents who have been intrigued by the possibility of
developing greater power at the expense of other institutions. Many across the ideological
spectrum continue to make use of decree laws that by-pass the legislature, and, many, regardless
of ideology, have considered Constitutional amendments that would allow them to extend their
Presidencies. Nonetheless, the recent efforts of leftist leaders particularly in Venezuela, Ecuador,
and Bolivia to widen Presidential power so that they are essentially no countervailing checks and
balances, and the support they received from their fellow citizens to do so, is a concern for those
concerned with the consolidation of representative democracy and democratic practices in the
region.
The concern with populist political leadership is not limited to those already in power.
Other potential leaders whose commitment to political pluralism may be in doubt wait in the
wings. Some have lost elections in the recent past. Coup-plotter former Lieutenant Colonel
Ollante Humala won a plurality of the votes in the 2006 first round Peruvian elections, but was
defeated in the run-off. The leftist Presidential candidate in Mexico was defeated by the
narrowest of margins in the 2006 elections, and since then has refused to accept the legitimacy of
the electoral process. In Paraguay, the decades-long hegemony of the Colorado Party is being
challenged by suspended Roman Catholic bishop Msgr. Fernando Lugo who espouses a mixture
of leftist and populist rhetoric. The legitimacy of representative democracy, at least, seems
called into question by many leaders and potential leaders.
But, like the tango, it takes two for the dance to work. And it is the other side of the
partnership that concerns us here: the beliefs and practices of the citizens who are electing those
Presidents. Presidents are expanding their powers not only because they can, as other
institutions are weak, but because many of their citizens are supporting this expansion, sometimes
loudly and insistently. There appears to be within the region a growing level of citizen support
for increased Presidential power, at the expense of other branches of government, and increased
support for the dismantling of checks and balance critical to a democratic political system. This
phenomenon has been noted for some time, most notably by Guillermo ODonnell, and his term
delegative democracy, in which citizens grant their newly elected Presidents extraordinary
powers (O'Donnell 1995).
It is not hard to find concrete examples of citizens urging Presidents to take strong,
unilateral measures, ignoring courts and legislatures. When President Fujimori closed down the
Congress completely in 1992, insisting that its members were corrupt and many were tied to a
violent revolutionary movement, Sendero Luminoso, strong majorities of Peruvian citizens
supported his actions (Seligson and Carrin 2002; Carrin and Zrate 2007). In Venezuela, large
majorities of voters have repeatedly reelected President Chavez, even though he has neutered the
legislature and the courts.
There is a long tradition of populists leaders in Latin America, emerging first in Brazil in
the 1930s with the government of Getlio Vargas of Brazil and in Argentina in the 1940s with the
rule of Juan Domingo Pern. Other key populist figures include Lzaro Crdenas (Mexico), Jorge
Gaitn (Colombia) and Jos Velasco Ibarra (Ecuador). While the brand of populism Vargas and
Pern espoused was rooted in the right, drawing inspiration from Mussolini and European
fascism, there also have been populist leaders of the left, such as Juan Velasco Alvarados
4
military dictatorship in Peru from 1968-1975. The term populism is sometimes confused
merely with charismatic, personalistic leaders who appeal to a trans-class base (Taggart 2000).
The unifying characteristic of populist governments is not their left-right orientation. Rather, it is
a core belief that the institutions of classical liberal democracy, especially the legislature and the
courts, are anachronistic, inefficient and inconsistent with their own interpretation of the will of
the people.2 Populist leaders propose instead to listen to the people and to personally carry out
their will, while isolating the rejectionists usually running roughshod over fundamental
democratic guarantees of civil liberties, especially free expression and the right to due process.3
Are citizens willing to forego the institutions of liberal democracy, and imbue their
elected leaders with unchecked power, returning to an earlier epoch of authoritarianism? Are
those who favor greater executive dominance more on the left, coinciding with the rise of leftist
Presidents? If so, does this pose a threat to further democratic development in the region? Do the
trends reflect little more than a pendular shift in mood Stimson (1998; 2004) found periodically
occurs in the U.S., or are they a sea change, that might ultimately represent a threat to democratic
consolidation?
The 2006-2007 AmericasBarometer surveys allows us, for the first time, to analyze these
questions in a unique way that goes beyond general impressions. We will first examine citizen
ideology, defined in terms of their self-identification with the political left or right, and then
examine whether Latin American voters actually use their political orientation as the basis of
their vote. We will also see whether their orientation is linked to their support of democratic
values, including institutional checks-and-balances and political tolerance.
Second, we will examine in more detail the characteristics of those who seem to hold
democratic values in low regard. Who are they as a group, and how do they differ from those
who value democratic values? Finally, based on this analysis, we will look at whether political
orientation, as well as those who do not esteem democratic values, are likely to pose a threat to
further democratic consolidation in the region.
The Political Orientation of Latin Americans is Important to Democracy
How far left are Latin American and Caribbean citizens, and is there a trend toward the
left? To the surprise of many, it turns out that ideologically they are actually slightly to the right
of most respondents world-wide. We are able to say this because for many years, beginning with
the Eurobarometer and the early World Values Surveys (WVS), public opinion questionnaires
around the world have included a 10-point left-right ideology scale, with a score of 1 selected by
those who self-define themselves on the far left of the political spectrum, and a score of 10 by
2
In this sense, Weyland (2001 14) best formally defines populism as: a political strategy through which a
personalistic leader seeks or exercises government power based on direct, unmediated uninstitutionalized support
from large numbers of mostly unorganized followers. This direct, quasi-personal relationship bypasses established
intermediary organizations or deinstitutionalizes and subordinates them to the leader's personal will.
3
Among the classic works exploring the elements of the populist tradition are Conniff (1999) and Malloy (1977). For
a modern take, see Roberts (forthcoming) .
5
those who place themselves on the far right. The arithmetic mid-point on this scale would be 5.5,
but since whole numbers are accepted as the only possible choices, focus group studies show that
5 is nearly universally viewed as the neutral (neither left nor right) point by respondents.4 It
turns out that most countries hover near the center of the scale, although there are exceptions,
with, for example Belarus in 1990 averaging 3.88 while Bangladesh in 2000 averaged 7.56. The
average score for the world, however, was 5.56 for the pooled WVS data of 84 countries,
incorporating over 267,000 interviews for all nations from the entire series from 1981 to 2004.
This indicates a slight right-bias world-wide.5 Unfortunately, the coverage of Latin America and
the Caribbean in the WVS has been limited, but of the four countries included in the most recent
round (2000-01) the following average values are observed: Chile, 5.22; Mexico, 6.55; Peru,
5.69; and Venezuela 6.32. In sum, world-wide, and in the limited set of countries from Latin
America using the WVS, on a 1-10 scale, averages are just slightly right of center.
The AmericasBarometer data for 2006-07, which included 20 countries in the region,
finds that the regional average is 5.71, very close to the world average reported above, although
slightly more to the right.6 Overall, therefore, the data reveal that the citizens of Latin American
countries have a slightly right political orientation.
Over time, however, The AmericasBarometer data also show a shift to the left over the
past two years, as we can see by comparing the 2004 surveys with the 2006-2007 surveys in the
countries that were surveyed both years. In 2004, the mean was 6.17; by 2006-2007, the mean
had dropped to 5.71. Most countries moved to the left, six of them significantly; one significantly
moved to the right, with the rest statistically unchanged. Thus, a slight shift to the left has
indeed occurred, and the trend is region-wide, but the magnitude of the shift is small and the
center of gravity still remains somewhat to the right. Moreover, longer-term trend data would
need to be examined before we could be confident that the move to the left in Latin America is
more than a transitory change.
Political Orientation Matters: People Vote their Beliefs
The question of whether Latin Americans vote on the basis of their political orientation is
an important one. In terms of the topic of this paper, whether the turn to the left presages a
decrease in support for democratic consolidation, we must link orientation to actual support for
4
A scale ranging from 0-10 would have provided a true neutral point, but the AmericasBarometer conforms to
world-wide standard of 1-10.
5
The WVS has expanded its range of countries over the years, moving from a concentration on advanced industrial
democracies to one that now includes many countries from the developing world. Looking exclusively at the 70
countries surveyed since 1999, the mean ideology score is 5.58, nearly identical to the entire series since 1981,
indicating no world-wide shift in the post Cold War epoch. World-wide, non-response on this question is typically
higher than on other survey items. The WVS mean is based upon 193,531 individuals who responded to the ideology
question on at least one wave of the WVS. In the AmericasBarometer, approximately 20% non-response was
encountered, which is typical for many surveys.
6
This comparison includes a subset of 10 countries from 2004 that were also surveyed with the identical survey item
in 2006.
6
the left in elections. For ideology to matter politically, it needs to translate into behavior
consistent with that orientation. If it does not, then there is little reason to be concerned about the
significance of shifts in ideological orientation. Since the survey not only included questions on
ideology but on party preference, we can analyze this relationship, and further examine results by
looking at social class, employment, and other factors that might cause societal cleavages.
However, the question of whether voters chose their leaders on the basis of political
orientation is important for other reasons. Within the study of political parties in Latin America,
researchers have grappled with this over time. Some have argued that most parties are highly
personalistic, and that voters choose their Presidents on the basis of their personal feelings
about him/her, rather than on the basis of the left/right orientation of the party from which he/she
is running (Gunther, Montero and Linz 2002). There is another long-standing literature on the
patrimonial party, that argues most people support a party based on their personal clientelist
linkages to the leadership, rather than on ideology and political beliefs (Linz 2000). These
competing explanations cannot be fully tested in this research, but we can explore in depth the
questions related to political ideology.
The survey research reveals not only the political orientation of voters, but also what party
they supported in the elections, and their personal characteristics (age, gender, income, region,
etc.) This allows us to get the most detailed empirical basis on the relationship of orientation to
vote, and the characteristics of the voters by party and political orientation.
In terms of poetical orientation, the data show that ideological dispositions along the
classic left-right continuum do indeed have a meaningful impact on how people vote. People
with more leftist orientations vote more for political parties on the left, and people with more
rightist orientations vote for parties on the right. However, the effect of ideology is different in
each country, as the party systems vary greatly. In more established democracies with
consolidate party systems, left and right parties are not very far apart ideologically, and their
voters are, similarly, drawn to the center, while in systems with deeper cleavages, the party vote
reveals a great polarization among the electorate.
This variation across countries can be illustrated through comparisons of some of the
countries in the sample of AmericasBarometer data. First, consider the case of Costa Rica,
compared to Nicaragua and El Salvador. In Costa Rica, Latin Americas most consolidated
democracy, ideological differences among the principal political parties around a left/right
continuum are minimal. In contrast, Nicaraguan parties show a significant ideological split and
parties in El Salvador are the most polarized of the three. In Figures I-1 and I-2 the mean
ideology score of those who supported particular presidential candidates are indicated with a
small circle.7 In Costa Rica, the average ideology score on the left-right scale in 2006-2007 was
5.9, and as can be seen in Figure I-1. All of the candidates who received a significant number of
votes in the election lined up very closely to that mean. (The greatest deviation was only .6 of a
point, and that was for voters supporting the Libertarian Party, at the very fringe of the Costa
Rican political spectrum.) The traditional right-of-center party, the PUSC, coincides precisely
7
The 95% confidence interval around that mean is shown by a horizontally placed I in the figures, such that the
larger the number of respondents who selected that candidate in the survey, the narrower the confidence interval.
7
with the national mean, whereas supporters of the traditionally left-of-center PLN, averaged 6.3,
slightly to the right of the national mean. These findings, incidentally, dramatically illustrate the
electoral realignment taking place in that country (Lehoucq 2006), as the PLN, the party of the
Nobel Peace Prize winner Oscar Arias, actually received support from voters slightly to the right
of the traditional center-right party. But the larger point is to note the very narrow range of
ideological difference in Latin Americas oldest and most stable democracy. Even though the
electoral scene has been marked with declining voter participation and evidence of declining
support for the system (Seligson 2002b), Costa Rica remains at the top of all of the countries in
the region in terms of political legitimacy, and, as these results show, the ideological
disagreements are very limited (Booth and Seligson 2005).
National ideology mean = 5.9
Oscar Arias (PLN)
Oscar Arias (PLN)
6.3
Ricardo Jaime Toledo (PUSC)

Presidential vote
Ricardo Jaime Toledo (PUSC)

5.9
Otton Sols (PAC)
Otton Sols (PAC)
5.7
Otto Guevara (ML)
Otto Guevara (ML)

5.3
1 2 3 4 5 6 7 8 9 10
Left Ideology Right

Source: AmericasBarometer by LAPOP
Horizontal lines indicate 95% confidence interval

Figure I-1. Ideological disposition and presidential vote choice: Costa Rica
8
Nicar agua
Dan iel Orteg a, FS LN
D aniel Ortega, FSLN
4.3
Pres i de ntia l v ot e
6.0
Enrique Bolaos, P LC
E nrique Bo lao s, PL C
S chaf ik H nda l (FM LN)

El Salvador
Schaf ik Hndal ( FMLN)
3. 3
Anton io S aca ( AREN A)
Antonio Saca (A RENA)

7.5
1 2 3 4 5 6 7 8 9 10
Le ft I deo lo gy Ri gh t
Ho riz on tal lin es i nd ica te 95 % co nfide nc e int erva ls
Figure I-2. Ideological disposition and presidential vote choice: Nicaragua and El Salvador
Consider now the more divided picture in two of Costa Ricas neighbors in Central
America, Nicaragua and El Salvador, depicted in Figure I-2. In Nicaragua, supporters of
Sandinista candidate Daniel Ortega are, as would be expected, considerably to the left of those
who supported Enrique Bolaos of the PLC. More importantly, however, the ideological gap
between left and right in Nicaragua, looking at the political orientation of the voters, is far wider
than in Costa Rica. In Nicaragua, therefore, partisanship is more determined by political
orientation than in Costa Rica, and there is much greater polarization of the political system
within the party system. This phenomenon is even more acute in El Salvador, where the
ideological distance is far greater still. Supporters of (the now deceased) Schafik Hndal of the
FMLN, the leftist party that emerged from the guerilla organizations of the civil war of that broke
out in the 1980s, averaged 3.3 on the scale, much more to the left than their left counterparts in
Costa Rica, compared to the 7.5 for supporters of Antonio Saca, head of the rightist ARENA
party that has won the presidency in every election since democracy was restored to that country.
Chile offers a final interesting case, with party and voters very closely aligned. As shown
in Figure I-3, the winning presidential candidate, socialist Michelle Bachellet, attracted voters
who were closest to the national ideological mean. Far to her left was Toms Hirsch, who
espoused a more radical program during the campaign, but won only a small portion of the votes
(thus explaining the wide confidence interval around the mean of the survey respondents who say
9
that they voted for that party). The rightist candidate, Joqun Levin, attracted voters who were
ideologically furthest to the right, while the center-right Sebastin Piera, who lost in the run-off
election held in January 2006, was supported by voters that the AmericasBarometer data show
were not as far right. In short, the voters in Chile hold ideologies that map perfectly onto the
spectrum of candidates from which they had to choose.
There are three key conclusions to draw from this review of ideology in Latin America
and the Caribbean. First, the median voter is slightly to the right of world opinion, even as more
leftist Presidents are elected in the region. Second, within that slightly rightist orientation,
voters moved somewhat toward the left, even in the brief span of years analyzed here 2004-2006.
Third, political orientation is an important determinant of voter preference in Latin America, with
its impact varying sharply by country.
Toms Hirsch
Toms Hirsch
2.6
Presidential vote
Joaqun Lavn
Joaqun Lavn
7.1
Sebastin Piera
Sebastin Piera
6.8
Michelle Bachellet
Michelle Bachellet
4.4
1 2 3 4 5 6 7 8 9 10
Left Ideology Right

Horizontal lines indicate 95% confidence intervals
Figure I-3. Ideological disposition and vote choice: Chile
Leftists are Less Likely to Favor some Core Democracy Values
Beyond the ballot box, the AmericasBarometer data show that ideology matters for the far
deeper and ultimately more profound issue of support for democracy. People who self-define as
more leftist tend to view their political systems as less legitimate and are less likely to favor
democracy as a political system less than do others. In some ways this is a very surprising
finding. Many of the leftists now in power came into politics fighting repressive regimes. In
10
many countries it was leftists who led the creation of new, more democratic constitutions and
fought for free and fair elections. An entire generation of leftists led the fight for democracy.
Yet, it is also true that historically, in the Latin American region as a whole, the
institutional trappings of democracy courts, legislatures, local governments have been
controlled by a small elite for its own benefit. The political systems of Latin America supported
economic policies that led to some of the highest income inequality in the world. It might not be
surprising that citizens wary of corrupt legislatures and non-transparent governments become
highly skeptical of institutional solutions to democracy. Their institutions have not served
them well.
How did we arrive at this conclusion that citizens who are more leftists are also less
likely to favor democratic values? This section will first analyze overall support for democratic
values. It will then link it to the left/right political orientation of citizens. Finally, it will add a
new dimension, a set of questions designed specifically to probe at support for executive
dominance, looking at both overall findings and breaking the findings down by political
orientation.
We used a number of methods. First, we developed a scale based on a question often

used to measure democratic support in contemporary democracy surveys, drawn from the work
in post-communist Europe of Mishler and Rose (1999 81-82), which has become known as the
Churchill question. Respondents are asked, on a 1-7 scale, whether they agree or not with the
statement: Even though democracy has many problems, it is better than any other form of
government. The scale was converted to the more familiar 1-10 range to match the second
item. This second item involved the use of a composite index to measure legitimacy. This index
is based upon the classic conceptualization of Lipset (Lipset 1959b) and Easton (Easton 1975).8 It
is comprised of five items that seek to tap the extent to which citizens trust their political system.
Both of these measures led to the same conclusion, as shown in Figure I-4: the more leftist the
respondents political orientation, the lower his belief that democracy is better than any other
form of government and the. These results are for the AmericasBarometer sample as a whole; in
one country, however, Chile, the pattern is reversed.
This is a disturbing finding, when coupled with the growing leftist political orientation of
citizens, and the salience of it in determining what Presidential candidate to support. For the
region as a whole, the finding that citizens of the Americas on the left are less likely to prefer
democracy than those on the right suggest that the movement to the left may represent a move
away from democratic values, while the legitimacy scale suggest that more citizens will question
the right of a democratic regime to govern.
In the AmericasBarometer survey respondents were also asked: There are people who
say that what we need is a strong leader who does not have to be elected via the ballot. Others
say that even though things dont work, electoral democracy, that is, the popular vote, is always
the best. Overall in the region, in 2006 only 14,6% percent or respondents preferred a strong
8
This is measured by a 5-item series, each scored on a 1-7 and then transformed into a 1-10 index. Details can be
found in (Seligson 2002a; Booth and Seligson 2005; Seligson 2006).
11
leader, a very encouraging sign for democracy. Yet, in 11 of the 15 countries where this question
was asked, a those who preferred a strong leader were more likely to be further to the left than
those who preferred electoral democracy. Only in Guatemala, Chile and the Dominican
Republic does the left more heavily support electoral democracy than the right, and reject a
strong leader. In Mexico, it was a tie.
However, by another standard measure of democratic values in the surveys, political

tolerance, this difference between left and right disappears, or, in some cases, has the reverse
tendency. In most countries, the left and right are equally tolerant. In three countries, Mexico,
Guatemala, and Chile, the left is more tolerant than the right. Interestingly, in prior studies of
Latin America conducted by LAPOP, the left has often been found to be more tolerant than the
right. The questions used to measure tolerance have centered on the willingness of respondents
to grant to opposition minorities basic civil liberties, such as the right to vote, run for office,
protest and enjoy free speech. It is possible that because the left is in power now in much of
region, that they have become less tolerant of the right, now out of power. This hypothesis,
however, would require considerable effort to test and prove, a subject beyond the scope of this
short chapter.
5.8
Legitimacy
5.6
5.4
Democracy is the
best system
5.2
5.0
4.8
1 2 3 4 5 6 7 8 9 10
Left Ideology Right

Figure I-4. Impact of ideology on legitimacy and belief that democracy is the best system
In sum, the left and right in Latin America about equally support one important
democratic value: political tolerance. However, they are sharply divided on whether democracy
12
is the best system of government, and in their willingness to support a non-elected leader, with
the left more inclined to favor a non-elected leader, and showing much greater skepticism for
democracy as a political system. Given at least a slight move to the left over the past two years
(although overall still generally rightist), and the increasing number of political systems headed
by leftists parties, if this trend continues, more and more Latin Americans may opt for non-
democratic alternatives when given the opportunity. There is great variance among countries,
however.
Latin Americans are not Ready to Jettison Democratic Institutions
While there are important differences between the left and right in their support of checks
on executive power, it is important to put this in context, looking at overall patterns of support
within the region. We have already noted the fact that in comparison with the rest of the world
Latin Americans remain slightly to the right of the mean on the left/right continuum of political
orientation. In this section, we will examine attitudes towards specific political institutions, to
better gauge where Latin Americans are likely to put their trust. We will also examine other
characteristics of Latin Americans, to see whether other traits, seem related to their support of
democratic institutions and checks and balances, or to put it another way, whether there are
certain groups of people more inclined to jettison these institutions.
Is there evidence that citizens of Latin America and the Caribbean would prefer populist-
style governments rather than liberal democracies? One way to measure this is to look at trust in
institutions. Nearly all surveys of Latin America and the Caribbean have found that citizens hold
legislatures and judiciaries in low regard. An examination of the AmericasBarometer data
confirms that picture. Respondents were read a long list of institutions and asked how much they
trust each one. Converted to the familiar 0-100 scale, the Catholic Church9 always scores at the
top, with a mean of 68 for the region as a whole. At the very bottom were political parties, with a
trust score of half that of the Church (35). The justice system averaged 44; the legislature and
supreme court, 45; and the police, 46. The armed forces, in contrast, scored second highest at
60, trailing the Church by only 8 points. In sum, of the state institutions covered in the surveys,
only the armed forces averaged on the positive end of the 0-100 continuum, while the key
representative institutions of liberal democracy were allon the low end.
How low are these scores compared to other countries in the world? The WVS, using a
slightly different measurement scale, allows us to roughly compare these scores to countries
outside the region. In the WVS around the world, the same pattern of trust prevails. An
examination of the WVS rank-order of the church, the military, the police, parties, the legislature
and the judiciary shows that the lowest average trust score for all countries world-wide is for
parties, followed by the legislature and then the justice system. In that sense, then, Latin America
conforms precisely to the international pattern. The AmericasBarometer findings provides even
more interesting comparisons, as it includes the United States and Canada in the data set,
allowing us to compare the Latin American and Caribbean results with identically worded and
9
In all countries except Jamaica and Guyana, respondents were asked about the Catholic Church, whereas in those
two countries they were asked about the Church.
13
scaled questions for the U.S. and Canada. Confidence in parties is low in Canada (49.1) and the
U.S. (42.9), but even so, in comparative terms Canada ranks higher than any country in the
region, while the U.S. ranks fourth. Only three other countries average above 40 on the 100-point
scale, and two, Paraguay and Ecuador average below 30. When it comes to the courts and the
supreme court, the gap between Canada and the U.S., on the one hand, with Latin America on the
other is greater. Canada scores 71.3 and the U.S. 67.1 for the supreme court, while all of the
countries in Latin America and the Caribbean scoring below 60 and half of them below 50. At
the very low end are Haiti (31.4), Paraguay (30.2) and Ecuador (24.7). For the system of justice,
the pattern is the same, with the U.S. and Canada the only counties to score above 60, and the tail
end being comprised of Peru (32.6) Paraguay (31.0) and Ecuador (28.0). It is of note that
Uruguay, among the best established democracy in the region, is the country closest to the U.S.
and Canada on this question, scoring 55.6, followed by Costa Rica (52.9), the oldest established
democracy in the region.
Given these low scores for the traditional institutions of liberal democracy, one might
expect populist sentiment to be high and citizens readily willing to jettison them. Yet, this is not
an accurate reading of the survey evidence. Despite the low support for parties, when asked one
of the standard items included in many democracy surveys: In your opinion, can there be
democracy without political parties? the AmericasBarometer 2006-2007 finds only a minority,
43%, of them agreeing. This is not uniformly true, however. In Ecuador, j 50.5% of respondents
agreed, while in Haiti 62.2% agreed that there can be democracy without parties. It is difficult to
determine if this is good news or bad; while in most countries a majority rejects the notion that
there can be democracy without parties, yet a strong minority accepts it. It is important to dig a
bit deeper to better interpret the implications of these findings for the proclivity of Latin
Americans to accept populist rule.
To gain further insight into the appeals of populism, LAPOP developed a new set of items
for the 2006-2007 AmericasBarometer round specifically designed to gauge the extent to which
citizens are willing to dispense with parties, legislative and judicial institutions, ceding power to
the executive. This provides the most compelling data to date on attitudes towards representative
institutions. Five items were constructed that formed a single dimension (using factor analysis)
when the data were analyzed. One item, for example, read: With which of these two opinions
do you agree more: 1) For the progress of the country, its is necessary that our presidents limit the
voice and vote of the opposition parties; 2) There is no reason that would justify that our
presidents limit the voice and vote of the opposition parties, even if they hold back the progress
of the country.10 An overall scale of support for populism was created, in which it was found
10
The other items in the series (with appropriate wording change for parlimentary vs. presidential systems) are:
POP2: 1. The Congress hinders the work of our presidents/prime ministers, and should be ignored 2. Even when it
hinders the work of the president/prime minister, our presidents/prime ministers shold not bypass the Congress.
POP3: 1. Judges frequently hinder the work of our presidents/prime ministers, and they should be ignored. 2. Even
when judges sometimes hinder the work of our presidents/prime ministers, their decisions should always be obeyed.
POP4: 1. Our presidents/prime ministers should have the necessary power so that they can act in the national
interest. 2. The power of our presidents/prime ministers should be limited so that they do not endanger our liberties.
POP5. 1. Our presidents/prime ministers should do what the people want even when laws prevent them from doing
so. 2. Our presidents/prime ministers should obey the laws even when the people dont want them to.
14
that one-third (32.6%) of the respondents for the sample as a whole refused to accept any populist
measure, and only 15.9% of the respondents would support more than two of the five populist
measures. Yet, 47.6% of the respondents in the pooled data set were willing to accept two of the
five measures, and 63.4% were willing to accept at least one such measure. By this accounting,
then, while only a small minority of citizens in Latin America and the Caribbean favor a wide
variety of measure to strengthening of the presidency at the expense of representative, liberal
democratic institutions, a substantial majority would accept at least some reductions in the
principle of separation of powers.
What of the minority that rejects liberal democracy on our Populism questions? In
democracies, minorities can be important, especially if they are concentrated in a homogeneous
sector of the population. In elections, for example, minorities with a unified position can achieve
electoral victory when the opposition is divided. Minorities also can carry great weight in the
street, should they decide to embark upon demonstrations, civil disobedience or engage in
terrorism. It is important to know, therefore where minority support for populist rule is strong in
Latin American and Caribbean area.
The results of our analysis echo the findings in the classical work by Seymour Lipset
(1959a), who expounded on the concept of working class authoritarianism that is, it is the
working class, rather than the elite, who exhibit more authoritarian tendencies.
Consistent with Lipsets work, regression analysis on the pooled data finds that populist
sentiment is significantly higher among the poorer and less well educated.11 The Lipset-
confirming results are depicted in Figures I-5 and I-6. A surprising finding, however, is that even
when controlled for wealth and education, the younger the age of the respondent, the more likely
s/he would be to support populist measures. This comes as a surprise to many, since it has long
been assumed that older people are set in their ways and therefore more likely to support a
government that controls dissenters. In fact, it is the youth of Latin America and the Caribbean
who are more likely to support populist measures. One needs to probe the data more deeply for
an explanation of these findings, but it is the case that many older citizens of the region
personally experienced the dictatorial military regimes that predominated in the 1970s. That
experience may have helped immunize them against the appeals of populist rule. They have
been there; they know the costs. The young, in contrast, know only the contemporary democratic
period, which has been filled with bitter disappointments for many; economic growth in much of
Latin America and the Caribbean has been modest, but more than that, it largely has not trickled
down to the poor. Moreover, it may well be that the young are the most susceptible to populist
appeals, while older citizens are more jaded, having seen politicians of all stripes come and go,
they may be resistant to the latest fad. Perhaps it is not surprising, therefore, that Chile and Haiti
are the two countries that show the strongest impact of age on rejection of populists.
The public opinion data presented here, therefore should not come as a surprise. The
implications of these results are potentially sobering; as the youth of today grow to become the
11
Analysis of individual country data, not reported here, finds variation in these regression patterns. Extensive
studies on each of the AmericasBarometer countries can be downloaded at www.lapopsurveys.org.
15
majority of voters, the findings here suggest increasing susceptibility of populist appeals.12 Of
course, as the young age, they may become more conservative and less willing to accept such
appeals, but in fact, the populist questions are not focused on a liberal/conservative dimension,
but instead are tapping into populist vs. democratic notions of governance. For those, one could
have expected that it would be the youth who would be less inclined to accept anti-democratic
measures.
34
32
32
Populism scale
30
Populism scale
30
28
28
25
26
22
24
None Primary Secondary University

0 1 2 3 4 5 6 7 8 9
Education
Wealth index
Figure I-5. Impact of education on support for Figure I-6. Impact of wealth on support for populism
populism
Conclusions: Political Institutions and Values Remain Important
The survey data presented here do indeed suggest that more leftist oriented voters are
somewhat more likely to jettison the political checks and balances needed in a democracy, and to
grant their elected executives too much power, leaving the door open for some form of electoral
authoritarianism. Given the rise of leftist Presidents, supported by these voters, there is ample
reason to be watchful. Many of the moderate leftist Presidents and their cadres may be
completely committed to and respectful of democratic institutions. This analysis has shown,
however, that at least some of their partisan supporters are less so.
Yet, there is at this point no immediate threat that the increasing support of the left will
pose a challenge to the consolidation of democracy. In fact, this research has uncovered some
underlying dynamics that are positive in terms of democratic institution-building. First,
supporters on the left remain in the political system and use the vote, rather than violence, to
make their political concerns felt. Second, the political parties in a number of countries are at
least mildly programmatic, reflecting the left/right continuum of the voters. Third, while there is
12
An obvious alternative explanation is that as the youth age they will come to resemble the older population and
thefore will be equally resistant to populist appeals. Unfortunately, the panel data to unpack age vs. cohort effects do
not exist (Firebaugh 1997).
16
little trust in most representative political institutions, citizens would like to keep them even
political parties, which are at the bottom on the heap on trust measures.
Yet, the data point to some overall trends that merit careful attention. The ideological
center of gravity in Latin America is, by world standards, slightly to the right, yet it attitudes are
moving to the left. Ideological cleavages in Latin America, long after the end of the Cold War,
in some countries still line up along a distinct left-right dimension, and voters support parties
consistent with their ideological orientations. The gap between left and right is very narrow,
however, in other countries (e.g., Costa Rica) but strikingly wide in other countries (e.g.
Nicaragua, El Salvador and Chile). Being on the left in Latin America and the Caribbean has
implications beyond the ballot box; for the region as a whole, those on the left are far less likely
to believe that their political system is legitimate and are less likely to believe that despite all its
flaws, democracy is still the best system. Moreover, in a number of countries, the left is more
likely to support strong leaders who do not need to be elected via the ballot. A second, related,
trend is the rise of populist figures and governments, especially in South America. This trend
emerges, no doubt, from the very low level of trust that many citizens hold of key institutions of
liberal democracy, especially parties, the courts and the legislature. In some countries, such as
Ecuador, trust in these institutions is at extremely low levels. This should come as no surprise
given the events in Ecuador over the past decade, in which a succession of democratically elected
presidents have been forced to leave office, while the executive has eviscerated any notion of
judicial independence. While demands for across-the-board measures that would essentially
create dictatorial rule is still very much a minoritarian sentiment for the region as a whole, close
to two-thirds of respondents in the region would accept at least some reduction in the institutional
autonomy of the courts, the legislature and opposition parties. Support for those measure that
would result in what Diamond has called a hollowing out of democracy (Diamond 1999), is
closely associated, in the region as a whole, with the poor, the less-well educated and the young.
Finally, there are other long-term trends of increasing inclusion at work as well, that may
work to strengthen support of democracy. It may well be that these findings are related to a
change in the long-term pattern of historical exclusion of the poor and ethnic minorities in the
region. The survey data for Bolivia, for example, show that soon after Evo Morales, whose roots
are in the indigenous community, was elected president of that country, the gap between
indigenous and non-indigenous support for democracy disappeared. Thus, the lower support for
democracy among those on the left could well shift as governments take power that are more
responsive to the demands of the excluded populations.
17
References
Booth, John A., and Mitchell A. Seligson. "Political Legitimacy and Participation in Costa Rica:
Evidence of Arena Shopping." Political Research Quarterly 59, no. 4 (2005): 537-50.
Carrin, Julio F., and Patricia Zrate. The Political Culture of Democracy in Per: 2006. Edited
by Mitchell A. Seligson. Lima: Instituto de Estudios Peruanos, 2007.
Conniff, Michael L., ed. Populism in Latin America. Tuscaloosa: University of Alabama Press,
1999.
Diamond, Larry. Developing Democracy: Toward Consolidation. Baltimore: Johns Hopkins
University Press, 1999.
Easton, David. "A Re-Assessment of the Concept of Political Support." British Journal of
Political Science 5 (1975): 435-57.
Firebaugh, Glenn. Analyzing Repeated Surveys. Thousand Oaks, Calif.: Sage Publications, 1997.
Gunther, Richard, Jose R. Montero, and Juan J. Linz. Political parties : old concepts and new
challenges. Oxford ; New York: Oxford University Press, 2002.
Lehoucq, Fabrice. "Costa Rica: Paradise in Doubt." Journal of Democracy 16, no. 3 (2006): 140-
54.
Linz, Juan. "Democratic Political Parties: Recognizing Contradictory Principles and
Perceptions." Scandinavian Political Studies 23, no. 3 (2000): 252-64.
Lipset, Seymour Martin. "Democracy and Working-Class Authoritarianism." American
Sociological Review 24 (1959a): 482-502.
. "Some Social Requisites of Democracy: Economic Development and Political
Legitimacy." American Political Science Review 53 (1959b): 65-105.
Malloy, J. M. Authoritarianism and Corporatism in Latin America. Pittsburgh: University of
Pittsburgh Press, 1977.
Mishler, William, and Richard Rose. "Five Years After the Fall: Trajectories of Support for
Democracy in Post-Communist Europe." In Critical Citizens: Global Support for
Democratic Governance, edited by Pippa Norris, 78-99. Oxford: Oxford University Press,
1999.
O'Donnell, Guillermo. "Delegative Democracy." Journal of Democracy 5, no. 1 (1995): 55-69.
Roberts, Kenneth M. Changing Course: Parties, Populism, and Political Representation in Latin
America's Neoliberal Era. Cambridge: Cambridget University Press, forthcoming.
Seligson, Mitchell A. "The Impact of Corruption on Regime Legitimacy: A Comparative Study
of Four Latin American Countries." Journal of Politics 64 (2002a): 408-33.
. "Trouble in Paradise: The Impact of the Erosion of System Support in Costa Rica, 1978-
1999." Latin American Research Review 37, no. 1 (2002b): 160-85.
. "The Measurement and Impact of Corruption Victimization: Survey Evidence from Latin
America." World Development 34, no. 2 (2006): 381-404.
Seligson, Mitchell A., and Julio Carrin. "Political Support, Political Skepticism and Political
Stability in New Democracies: An Empirical Examination of Mass Support for Coups
D'Etat in Peru." Comparative Political Studies 35, no. 1 (2002): 58-82.
Stimson, James A. Public Opinion in America : Moods, Cycles, and Swings. 2nd ed,
Transforming American politics. Boulder, Colo: Westview Press, 1998.
. Tides of consent : How Public Opinion Shapes American Politics. New York:
Cambridge University Press, 2004.
18
Taggart, Paul A. Populism, Concepts in the Social Sciences. Buckingham [England] ;

Philadelphia: Open University Press, 2000.
Weyland, Kurt. "Clarifying a Contested Concept: Populism in the Study of Latin American
Politics." Comparative Politics 34, no. 1 (2001): 1-22.
19
II. Illiberal Democracy and Normative Democracy: How is

Democracy Defined in the Americas?
Julio F. Carrin
Abstract
How do Latin Americans define democracy? Following a methodology developed by the
AfroBarometer, this paper finds that the majority of them do so in normative terms, and that
those who do so are more likely to endorse it than those who define it in instrumental or negative
terms. But I also find that the effect of defintional differences is not very strong. The data show
that the majority of Latin Americans endorse democracy, but that there are important cross-
national variations. Moreover, the study finds that this support is rather shallow. Many of the
self-declared democrats hold views that are inimical to democratic rule, such as endorsement of
military coups or willingness to support illiberal exercises of political power. I find that
depending on the question, as few as 18 percent and as many as 47 percent of self-declared
democrats are ready to justify a military coup. I also find that a quarter of those who endorse
democracy have no problem agreeing with the statement that presidents should limit the voice
and vote of the opposition if it is necessary for the progress of this country. These findings thus
suggest very careful analysis is called for when using approaches to measuring democracy based
on respondent definitions of it.
An important milestone in the analysis of public attitudes towards democracy was the
publication by the United Nations Development Programme (UNDP) of Democracia en Amrica
Latina (PNUD-PRODDAL 2004). Th UNDP report offers a gloomy assessment of Latin
Americans commitment to democracy. Although it finds that a majority of the respondents
support democracy as an ideal form of government, it argues that this support is shallow.
According to the study (PNUD-PRODDAL 2004, 132), a significant segment of the public is
willing to sacrifice a democratic form government in exchange for real socioeconomic progress.
The UNDP report (PNUD-PRODDAL 2004, 134) classifies 43 percent of Latin Americans as
democrats and 26 percent of them as non-democrats. The UNDP identifies a category for
those who are ambivalent, that is, have contradictory views about democracy: they choose
democracy as their preferred system of government but are also willing to support authoritarian
or delegative acts when circumstances merit it. According to this report (PNUD-PRODDAL
2004, 137), what differentiates democrats from those who are ambivalent is their attitude toward
the tension between democracy and economic development: democrats are not willing to
sacrifice the former for the latter whereas the ambivalents are.
Based on the UNDP study, it appears that the commitment to democracy among many
Latin Americans is, to use the term developed by Bratton and Mattes (2001a), merely
Associate Professor at the University of Delaware in the US and Researcher at the Instituto de Estudios Peruanos.
21
instrumental. That is, people support democracy because of the material benefits expected from
it and not because of the normative values that democracy embodies. In their study of public
attitudes toward democracy in Africa, Bratton and Mattes (2001a) confront this issue of the
instrumental versus normative foundations of support for democracy head on, and find
contradictory evidence. On the one hand, they report (2001a, 473) that [t]he fact that African
survey respondents support democracy while being far from content with its concrete
achievements suggests a measure of intrinsic support for the democratic regime form that
supersedes instrumental considerations. On the other hand, they also argue (idem) that the
general public in African countries thinks instrumentally: in other words, support for democracy
hinges critically upon popular approval of government achievements.
Echoing this ambivalence about the foundations of the support for democracy, Bratton
and Mattes (2001b, 118-119) write in another article that in at least five of our six countries [in
Africa], popular support for democracy has a strong instrumental component. Citizens extend
support to a democratic regime in good part because they are satisfied with its performance in
delivering desired good and services. Yet 21 percent of all survey respondentssay that they
support democracy in principle even though they are dissatisfied with the performance of their
own regime. These citizens value democracy intrinsically, that is, not merely as a means of
delivering development but as an end in itself. A more recent study (Mattes and Bratton 2007),
confirms that support for democracy in Africa is driven by both normative concerns and
instrumental considerations.
Two conclusions can be drawn from the AfroBarometer results. First, when we lear than
that the majority of people support democracy in Latin America, we should conclude that this
finding is not the end of the analysis but the beginning of it. What kind of democracy do people
have in mind when they say they support democracy? What important cross-national variations
in these definitions and in the levels of support for democracy can be found in Latin America?
How deep and consistent is the commitment to democracy? What are the factors that foster a
stronger commitment to democracy? The second point to be made is that we do not need to ask
whether support for democracy in Latin America is driven by normative commitments or
instrumental considerations. Clearly, it is driven by both. What we need to do is unpack the
idea of instrumental support. For instance, do people accept democracy because they like the
incumbent, or perceive an improvement in their economic condition, or because they trust their
political institutions? Is it the perception of personal threat that drives their attitudes towards
regimes, or is it peoples satisfaction with the way democracy is working in general?
In the following pages I address these questions. I start the analysis by examining
peoples conceptualizations of democracy. Then I examine how these conceptualizations relate
to support for democracy. After that, I examine the depth of democratic commitments, exploring
mass support for liberal and illiberal forms of democracy. I propose a more robust way to
measure attitudinal support for democracy. I conclude the analysis by identifying the factors
associated with support for liberal democracy. A summary section ends the chapter.
22
Mass Conceptualizations of Democracy
Before proceeding with the analysis of the conceptualizations of democracy among the
public, it is important to examine first the overall degree of support for democracy in the region.
We measure this with the item that asks respondents if they prefer a democracy, a dictatorship or
are indifferent as to regime-type preference. For the entire Latin American and Caribbean sample,
67.1 percent prefer democracy while 14.0 percent support an authoritarian government, and 11.4
percent are indifferent (an additional 8.4 percent does not answer the question). When those who
dont answer the question are removed from the calculation, support for democracy rises to 73
percent in Latin America and the Caribbean.
In some countries, however, people are quite reluctant or unable to declare a regime
preference. In Guatemala almost 20 percent of the respondents refuse to answer the question. In
Mexico, Colombia, and Nicaragua, ten percent or more of the respondents are in similar situation
(Haiti is a borderline case, with 9.6 percent declining the answer the question). One could
hypothesize that these countries have large rural and/or poorly educated populations that do not
have the sophistication of expression to answer a question like this. This, however, does not
explain why refusal rates are much lower in countries such as Bolivia, Ecuador, or Honduras,
which also have large rural and poorly educated populations. Perhaps support for democracy is
related to the ability to define it, as Mattes and Bratton (2007) argue. This section explores this
hypothesis.
The 2006-2007 AmericasBarometer questionnaire asked the following open-ended

question: In a few words, what is the meaning of democracy for you? Respondents could
provide up to three different definitions. They were then asked to specify (if they mentioned more
than one definition) which of their definitions was the most important. Table II.1 reports the
summary of the multiple answers given to this question (the table does not include respondents
from Canada and the United States). Three quarters of Latin American and Caribbean
respondents were able provide a definition of democracy. Interestingly, this is almost the exact
proportion of respondents who could provide a definition of democracy in a study of six African
nations in 1999-2000 (Bratton and Mattes 2001b, 108). In Canada, in contrast, only 15.5% could
not provide any response, while in the U.S. it was only 8.5%. Among the different meanings
assigned to democracy, the single most frequent was freedom of expression (18.9 percent)
followed by liberty (10.4 percent).
Before grouping these varied definitions into more parsimonious clusters it is necessary to
examine cross-national variations in the ability to provide any definition of democracy. To
analyze cross-national variations, I collapse the information provided in Table II.1 into two
groups. The first group is composed of all the respondents who provided a definition of
democracy, even if it had a negative connotation. The other group was composed of all of those
who could not assign a meaning to the term. Figure II.1 gives the distribution of this variable
(ability to define democracy) in the countries in which this question was asked in 2006-2007.
23
Table II-1. Mass Conceptualizations of Democracy, 2006-2007
Frequency Percent Valid Percent

It does not have any meaning 6958 22.3 24.7
Liberty (without specifying what type) 3237 10.4 11.5
Economic liberty 457 1.5 1.6
Liberty of expression 5907 18.9 20.9
Liberty of movement 427 1.4 1.5
Liberty, lack of 165 .5 .6
Being independent 563 1.8 2.0
Well being, economic progress, growth 587 1.9 2.1
Lack of well being, no economic progress 106 .3 .4
Capitalism 35 .1 .1
Free trade, free business 128 .4 .5
Employment, more opportunities of 400 1.3 1.4
Lack of employment 210 .7 .7
Right to choose eaders 991 3.2 3.5
Elections, voting 493 1.6 1.7
Free elections 451 1.4 1.6
Fraudulent elections 20 .1 .1
Equality (without specifying) 1181 3.8 4.2
Economic equality, or equality of classes 303 1.0 1.1
Gender equality 189 .6 .7
Equality before the law 283 .9 1.0
Racial or ethnic equality 119 .4 .4
Equality, Lack of, inequality 103 .3 .4
Limitations of participation 68 .2 .2
Participation (without specifying what type) 572 1.8 2.0
Participation of minorities 82 .3 .3
Power of the people 574 1.8 2.0
Human rights, respect rights 963 3.1 3.4
Disorder, lack of justice, corruption 322 1.0 1.1
Justice 310 1.0 1.1
Obey the law, less corruption 219 .7 .8
Non military government 82 .3 .3
Live in peace, without war 830 2.7 2.9
War, invasions 21 .1 .1
Other answer 854 2.7 3.0
Total 28210 90.4 100.0
Missing 3000 9.6
Total 31209 100.0
24
100%
8.5 9.7
12.5 12.8
14.8 15.5
19.3 21.3 21.8 23.0 25.0
27.1
31.3 31.4 32.0 32.1 33.5 34.6 34.6
37.5
80%
No
Yes
60%
%
91.5 90.3
87.5 87.2 85.2 84.5
40% 80.7 78.7 78.2 77.0 75.0
72.9
68.7 68.6 68.0 67.9 66.5 65.4 65.4
62.5
20%
0%
U
Ve
Pe a
Pa o
Ja
Pa ica
C ua
N ma
G agu
D na
El R e
H lvad
B ura
SA
ru
os y
hi
a n e la
ai
ol
ic
om
on
ra
ua a
uy
ex
m
ne
ru
na
ra
Sa p.
ti
ar la
le c a
om
zi s
gu
ta
ad
ic
te
d or
a
.
zu
l
a
bi
i
a
Figure II-1. Ability to Define Democracy
Figure II.1 shows two clusters, one in which 80 percent or more of the respondents can
provide a definition of democracy (Peru, Canada, Chile, Costa Rica, Uruguay, Venezuela, and the
United States), and another where less than 70 percent of the interviewees can do so (Colombia,
Guatemala, Nicaragua, Dominican Republic, El Salvador, Guyana, Brazil, and Honduras). In the
first group, with the exception of Peru, we have countries with a strong tradition of democratic
rule (despite the 1973-1990 interregnum in Chile, and recent events in Venezuela). The second
group, with the exception of Colombia and Guyana, is composed by Central American countries.
Bratton and Mattes (2001b, 117) argue that individuals who cannot define democracy
are much less attached to it as a preferred form of regime. The overall evidence from the
Americas seems to support this view. When the entire pooled data (without including
respondents from the United States and Canada) are analyzed, the ability to define democracy is
found to increase support for it,1 and the effect is moderately strong. For instance, 65 percent of
those who cannot provide a definition of democracy are quite willing to say that they prefer it
1
For purposes of this analysis, non-support for democracy is defined as those who did not choose the option
democracy is always preferable. Those who did not answer the question on support for democracy were removed
from the analysis.
25
over authoritarianism.2 Among those who can define it, support for democracy rises to 75
percent. The difference is 10 percent points and although statistically significant, it is not
overwhelming.
In some countries in the Americas, however, the impact of the ability to define democracy
on its support is quite strong (Figure II.2). In the advanced democracies, we see a very strong
correlation between the ability to define democracy and support for it. In the United States, for
instance, those who can provide a definition of democracy were twice as likely to support it as
those who cannot define it. In Canada, the difference in the levels of support for democracy
between those who can define it and those who cant is about 30 percentage points. In Latin
America and the Caribbean, we also find that some countries exhibit a strong correlation between
the ability to define democracy and support for it. In Uruguay the difference is 32 points. In Peru
support for democracy is 19 percent points higher among those who can offer a definition than it
is among those who cannot. In Paraguay the difference is 27 percent points; in Colombia the
difference is 15 points. In other countries, such as Guatemala, Honduras, Panama, the Dominican
Republic, Haiti, Jamaica, and Costa Rica there is no statistically significant difference in the
levels of support for democracy between those who can define it and those who cannot. No
pattern is immediately clear to explain why in some countries the ability to define democracy is
so intertwined with its level of support.
2
Support for democracy is measured by agreement with the following option: Democracy is preferable to any other
form of government. Support for authoritarianism is measure by agreement with the following: Under some
circumstances an authoritarian government can be preferable to a democratic one. Regime indiference is measured
by agreement with the following: For people like us, a democratic regime is the same as a non-democratic one.
26
102%
96% Can't define

90% Can define
84%
78%
% Support Democracy
72%
66%
60%
54%
48%
89 88
86 85 86
42% 81
84
82 81 80
79 80 79
76 75 77
75 75 74
36% 71
69 70
72 71
66 66 66
65 64 64
62 61
30% 59 61 61
53 54
52
24%
40
18%
32
12%
6%
0%
P
C
am
ar
st
l
ol
ra
hi
o
ic
uy
ru
on
a

r
st
a
it
x
a
le
a
si
gu
a
an
e
a
te
d
ic
m
g
ra
a
m
zu
lv
u
l
ic
o
u
m
a
li
ra
R
g
a
a
e
ia
a
u
ic
d
y
la
U
la
a
or
a
n
D
id
om
o
s
in
ic
an
a
Figure II-2. Support for Democracy by the Ability to Define it
I nowreturn to the definitions of democracy listed in Table II.1. These can be grouped in
more parsimonious ways to facilitate analysis. The question is what categories to use?
Democracy is a concept with multiple meanings that is the product of more than 2,500 years of
history (Dahl 1989; Dunn 2005; Held 1996; Touraine 1997). There are many ways to classify the
multiplicity of its definitions. One is to group all the possible definitions into the traditional
meanings associated with democracy: liberty, equality, protection against arbitrary rule, and
participation. This is the strategy that, in general terms, was initially adopted by C. B.
Macpherson (1977) and later developed by Held (1996). Another way is to follow Dahls (1956)
insight, and define democracy in either procedural or substantive terms.
Yet another way of approaching this task is to define democracy based on the source of
rationality underscoring the belief (Bratton and Mattes 2001a; Mattes and Bratton 2007; Sarsfield
2003; Sarsfield and Echegaray 2006; Sarsfield and Carrin 2006). One could hold a given belief,
in this case democracy, because it is associated with certain values that are considered to be
desirable (i.e., liberty, equality, justice). Alternatively, one could hold that belief because it
produces certain goods or utilities (i.e., economic progress, development). In the first case, the
commitment to democracy is normative because it is based on an axiomatic rationality; in the
27
second case, it is instrumental because it is based on a means/end rationality. Democracy can

also be defined, from this perspective, in pejorative or negative terms, because it produces an
undesirable outcome (i.e., civil war, violence, disorder). And, of course, people could have no
conceptualization of democracy at all (an empty definition). This is the classificatory strategy
that it will be followed here because it builds on Bratton and Mattess successful and
pathbreaking analyses of African public opinion. Table II.2 lists the definitions presented in
Table II.1 and how they relate to our four definitions of democracy based on the rationality of the
belief.
We can now examine if the way people define democracy is related to its support. The
data show that those who define democracy in normative terms are only slightly more likely to
support it than those who define it in instrumental terms (77.2 and 71.9 percent, respectively).
The most important difference in the levels of support for democracy is found between those who
define it in negative terms and those who conceptualize it in normative terms: 64.4 percent and
77.6 percent, respectively (Figure II.3; data from the U.S. and Canada are not included in this
figure). In only a handful of countries, the difference in the level of support for democracy
between those who define it in normative terms and those who define it in instrumental terms
exceeds 10 percent points. These countries include the Dominican Republic, Mexico, and Chile.
In seven countries, the differences between those who see democracy in normative ways and
those who define it instrumentally are less than 5 percent points. Thus, while one can perceive an
association between certain definitions of democracy and support for it, one can conclude with a
certain degree of confidence that support for democracy in Latin America is not primarily
determined by the way people define it but rather by the ability to define it (in either positive or
negative ways).
28
Table II-2. Conceptualizations of Democracy Based on the Rationality of the Belief

Normative Instrumental Empty Pejorative
-Liberty (without specifying what type) -Free trade, free -It doesnt have -Lack of liberty
-Liberty of expression, voting, choice and human rights business any meaning -Lack of well
-Liberty of movement -Economic liberty -Other answer being, no
-Liberty, lack of -Well being, economic -DK/NA economic progress
-Right to choose leaders progress, growth -Lack of
-Elections, voting -Capitalism employment
-Free elections -Employment, more -Fraudulent
-Equality (without specifying) opportunities of elections
-Economic equality, equality of classes -Lack of equality,
-Gender equality inequality
-Equality before the laws -Limitations of
-Racial or ethnic equality participation
-Participation (without specifying which type) -Disorder, lack of
-Participation of minorities justice, corruption
-Power of the people -War, invasions
-Respect for human rights
-Justice
-Obey the law, less corruption
-Non-military government
-Live in peace, without war
80.0%
Pejorative
Empty
Instrumental
Normative
60.0%
%
40.0% 77.6%
71.9%
66.1%
64.4%
20.0%
35.6%
33.9%
28.1%
22.4%
0.0%
Doesn't Support Democracy Support Democracy
Figure II-3. Conceptualizations of and Support for Democracy
29
How Consistent is the Support for Democracy?
It should be said from the outset that an expressed preference for democracy does not
inoculate the Latin American public from embracing illiberal or even outright authoritarian
alternatives. The inconsistencies of mass preferences for democracy are not a novel finding. In a
pioneering and classic study, James Prothro and Charles Grigg (1960) found that people who
manifest support for democracy in the abstract do not always endorse democratic choice in more
concrete domains. What we need to ask is, then, how deep or consistent is expressed support for
democracy, and what factors make people more consistently supportive of democracy? To
answer this question I will, first, explore the extent to which self-professed democrats3 also
endorse authoritarian and illiberal choices. In the next section I explore what factors are
associated with preference for liberal forms of democratic rule.
A traditional threat to democracy in Latin America has been the sudden overthrow of
civilian governments by a military coup. One would expect that those who endorse democracy
over authoritarianism would be less inclined to support a potential coup dtat. This is indeed the
case. Figure II.4 reports the mean scores of support for military coups by regime preference. As
expected, the strongest support for military coups is found among those who declare a preference
for an authoritarian regime and the lowest is reported among those who declare a preference for
democracy. But the important point of this figure is that a sizable portion of self-professed
democrats are ready to support military coups under some circumstances.4 For instance, 17
percent of those who endorse democracy declare that a military coup to confront high
unemployment would be justified. Similarly, and more worrisome, 45 percent of self-declared
democrats would justify a military coup to fight delinquency. In a similar vein, 40 percent of
those who endorse democracy would justify a military coup to fight excessive corruption.
Clearly, determining the proportion of people who endorse democracy as an ideal form of
government is a necessary first step in our efforts to examine the diffusion of democratic attitudes
in the region. But the data just presented also suggest that to determine how deep democratic
attitudes are in the region, we need to look beyond the formal endorsement of democracy as an
ideal form of government.
3
Those who choose the option Democracy is preferable to any other form of government in the question described
in footnote 2 are considered self-professed democrats here.
4
Our survey asked a series of questions, including whether a military coup will be justified under a series of
circumstances: to confront high unemployment, to control social protest, to stop delinquency, to fight inflation, and
to fight corruption. For each positive answer, a score of one was assigned. The resulting index was then transformed
so it would range from zero (no support for coups under any circumstances) to 100 (support under all circumstances).
30
55%
50%
Mean Scale of Support for Coups

45%
40%
35%
30%
25%
43.4
20%
36.1
15% 29.2
10%
5%
0%
Democracy is always For people like us, it doesn't Authoritarian government
preferable matter could be preferable
Error bars: 95% CI
Figure II-4. Mean Scores of Support for Military Coups by Regime Preference
Today, the authoritarianism that Latin America faces is different from the traditional
military coup. In a disturbing development, some Latin American democracies seem to have
been at least partially undermined by elected presidents who use their considerable popular
support to erode the precarious democratic foundations of their respective nations. This situation
is a manifestation of a larger trend of democratic devaluation in the developing world, a
phenomenon characterized by Diamond (1999) as the globalization of hollow democracy.
While externally maintaining formal democratic institutions, many elected governments
systematically engage in anti-democratic behaviors aimed at the elimination of credible
challenges to their rule. The emergence and growth of these regimes have led some observers to
characterize them as electoral or competitive authoritarianism (Carrin 2006; Diamond 2002;
Levitsky and Way 2002; Schedler 2006). Democratic reversals in Latin America are more likely
to come today from publics who elect and reelect overbearing presidents bent on aggrandizing
their own powers at the expense of the legislature and the courts than from traditional military
coups.
Until recently, former President Alberto Fujimori of Peru was the best example of this
emerging authoritarianism, but the 2000 reelection of coup-plotter Hugo Chvez in Venezuela
along with subsequent developments there suggest the establishment of another electoral
authoritarian regime in the region. Recent political crises in Bolivia and Ecuador have at their
root the fears of the opposition that a similar path may be engulfing their countries, as presidents
use their popularity to undermine existing (and very unpopular) legislatures. In extreme cases,
elected presidents have openly assumed dictatorial powers, as was the case in Peru in April of
1992 and in Guatemala in May of 1993 (an attempt that was eventually defeated). Evidently no
single reason can explain why the presidential authoritarian takeover was successful in Peru and
31
not in Guatemala, but popular support/opposition was certainly an important factor in these
diverging outcomes. Whereas Fujimoris self-coup generated significant public approval in Peru
(Conaghan 1995; Carrin 2006), Serranos similar action was rejected outright in Guatemala
(Cameron 1998, 133). This discrepancy highlights the risks of mass support for civilian forms of
authoritarianism. As the PNUD-PRODDAL report concludes (2004, 137), democracies become
vulnerable when, among other factors, authoritarian political forces find in citizens attitudes a
fertile ground to act.
For this reason, the AmericasBarometer 2006-2007 probes potential mass support for
illiberal acts against the legislatures, the courts, the opposition, and the laws in general by elected
chief executives. In order to do that, the 2006-2007 questionnaire developed a new series, asking
the following questions:
I am going to read several pairs of statements. Taking into account the current situation of this country, I would like
you to tell me with which of the following two statements you agree with the most?
POP1. [Read the options]
1. It is necessary for the progress of this country that our presidents/prime ministers limit the voice and vote of
opposition parties [or, on the contrary]
2. Even if they slow the progress of this country, our presidents should not limit the voice and vote of opposition
parties.
1. The Congress hinders the work of our presidents/prime ministers, and should be ignored [or, on the contrary]
2. Even when it hinders the work of the president/prime minister, our presidents/prime ministers should not bypass
the Congress.
1. Judges frequently hinder the work of our presidents/prime ministers, and they should be ignored. [or, on the
contrary]
2. Even when judges sometimes hinder the work of our presidents/prime ministers, their decisions should always be
obeyed.
1. Our presidents/prime ministers should have the necessary power so that they can act in the national interest. [or,
on the contrary]
2. The power of our presidents/prime ministers should be limited so that they do not endanger our liberties
1. Our presidents/prime ministers should do what the people want even when laws prevent them from doing so. [or,
on the contrary]
2. Our presidents/prime ministers should obey the laws even when the people dont want them to.
It is encouraging to report that in all five questions, the liberal answer (presidents should
follow the law) prevailed over the illiberal response. Without including those who did not
volunteer an answer, 72 percent declared that presidents should not limit the opposition, 80
percent asserted that they should not bypass the Congress, 80 percent demanded that they obey
judges, 59 percent wanted their powers to be limited, and 70 percent thought that they should
obey the laws.5
5
All these five questions were added in a summary index of support for liberal rule. When a respondent answered
four of these questions, the average value of these questions was imputed to the missing fifth answer. This was done
in an effort to keep as many valid cases as possible. To facilitate the analysis, each question was numerically
transformed so that the resulting index will have a range from 0 to 100. Values below 50 indicate low support for
liberal rule, and consequently values at or above 50 indicate high support for liberal rule. For the entire sample, 82
32
Despite these encouraging figures, it is also clear that many self-professed democrats
chose the illiberal response in these questions. For instance, 25 percent of those who chose
democracy as the ideal form of government selected the illiberal answer in question POP1. In
POP4, 41 percent of self-professed democrats chose the illiberal answer. This, of course, raises
the issue of how deep or consistent is their democratic commitment.
One way to measure support for democracy in a more robust fashion is to combine regime
preferences with attitudes about how political authority should be exercised. If we take regime
preferences,6 collapsing the three choices into two, one for the democratic choice and the other
for those who prefer authoritarianism (or are indifferent) and combine them with the index of
support for liberal rule (dichotomizing the index into low and high support) we obtain the
following theoretical situations (Table II.3).
Table II-3. Conceptual Relationship between Regime Preference and Constraints on

Political Authority
Regime Preference
Constraints on political
authority Democracy Non Democracy
Liberal rule Liberal democracy Liberal authoritarianism
Illiberal rule Illiberal democracy Illiberal authoritarianism
The optimal choice for the consolidation of a healthy democracy is the combination of
attitudes that favor democratic rule with constraints on the power of presidents or prime ministers
to exert their political power. Those who exhibit such a combination of preferences are labeled
here liberal democrats. I will center the analysis on them. All the other choices are sub-
optimal for the prospects of democratic rule, but the worst possible combination is the one that
pairs preference for authoritarian forms of government (or indifference between them) with
support for the illiberal exercise of executive power.
The distribution of these attitudes for the entire sample is shown in Table II.4. While 73
percent of the respondents favor democracy (this figure excludes those who do not answer the
question), 61 percent of all respondents can be labeled as liberal democrats, that is, they support
democracy and constraints on the exercise of political power. It is encouraging to find that only 7
percent of those interviewed in 2006-2007 favor the worst possible combination of support for
illiberal rule and preference for non democracy.
percent of the respondents have values at or above 50 in the scale. These questions were not asked in Bolivia and
Ecuador.
6
Measured by the responses to the following question: With which of these phrases are you most in agreement?
Democracy is preferable to any other form of government; under some circumstances an authoritarian government
can be preferable to a democratic one; for people like us, a democratic regime is the same as a non-democratic one.
33
Table II-4. Support for Liberal Democracy and Illiberal Democracy

Support for Democracy
Democrats Non Democrats Total
Constraints Liberal Rule 60.6% 18.9% 79.5%
on power
Illiberal Rule 13.6% 6.9% 20.5%
Total 74.2% 25.8% 100.0%
Is this a valid or useful classification? One way to validate it is by using it to predict

certain attitudes that are related to support or rejection of democratic values. If the classification
has validity, it should help us predict, for instance, levels of political tolerance, with liberal
democrats being more tolerant that illiberal non democrats. Similarly, if this is a valid
classification it should be able to differentiate between those who endorse social authoritarian
values from those who do not, with liberal democrats being less socially authoritarian than
illiberal authoritarians.7 Figures II.5 and II.6 display the means of political tolerance and
authoritarian values for each our four regime x authority groups. In each case, the results are in
the expected direction and, more importantly, they are monotonic, meaning than liberal
democrats are more tolerant and less socially authoritarian than illiberal democrats, and in turn
illiberal democrats are more tolerant and less conservative than liberal authoritarians, with the
illiberal authoritarians ranking last in both attitudes.
7
It should be noted that the questions utilized to measure social authoritarianism do not have a political content.
They ask whether the respondent agrees with the following statements: a very effective way of correcting
employees mistakes is to criticize them in front of other employees; the person who contributes the most money to
the home is the one who should have the final word in household decisions; at school, children should ask
questions only when the teacher allows it; when children behave badly, parents are justified in occasionally giving
them a spanking.
34
65%
Mean Scale of Political Tolerance

60%
55%
60.9
55.7
50% 54.5
51.4
45%
Liberal Illiberal Liberal Illiberal
Democrats Democrats Authoritarians Authoritarians
Error bars: 95% CI
Figure II-5. Political Tolerance by Regime Preference and Support

for Liberal Rule
35%
Mean Scale of Authoritarian Values
30%
32.9
31.5
25%
29.5
27.1
20%
Liberal Democrats Illiberal Democrats Liberal Illiberal
Authoritarians Authoritarians
Error bars: 95% CI
Figure II-6. Social Authoritarianism by Regime Preference and

Support for Liberal Rule
35
As expected, there is important cross-national variation in the distribution of liberal

democrats (Figure II.7). Countries with strong democratic traditions, such as Costa Rica and
Uruguay, and those with strong parliamentary histories, such as Jamaica, report the highest
number of liberal democrats. Honduras, Venezuela, the Dominican Republic, Guyana, and Chile
follow. On the other hand, Haiti, that has the third highest percentage of people declaring a
preference democracy over authoritarianism, falls to next to last place when we qualify
preference for democracy with support for liberal rule. Paraguay exhibits the lowest levels of
attitudinal support for liberal democracy. Guatemala, Brazil, Nicaragua, and Peru, all countries
with spotty democratic records, follow Paraguay and Haiti as having the lowest percentages of
liberal democrats.
To have a better understanding of the factors that help discriminate between liberal
democrats and those who have other regime/authority preferences it is necessary to go beyond
national differences. It is imperative to explore the personal characteristics of each of these
groups. This is what it is done in the next section.
84%
81%
78%
75%
72%
69%
66%
Mean Support for Liberal Democracy
63%
60%
57%
54%
51%
48%
45%
42%
39%
74.8 73.9
36% 71.9 71.3
33% 67.0 65.6 64.0
30% 61.4
59.3 58.7 58.3 57.3
27% 55.3
51.1 50.8 50.0
24% 49.0
21% 42.7
18%
15%
12%
9%
6%
3%
0%
U
Ja Ric
H ca
Ve ra
G Rep
El o
Pa vad
Pe bia
H ma
Pa
ru
os y
on
om ela
hi
ol
ic
ra ua
ai
uy
ua
ex
m
ne s
Sa
na or
ru
ra
ar
ti la
le
om
zi
gu
ta
du
an .
ic
te
ai
gu
.
zu
m
l
l
ag
a
ay
a
Error bars: 95% CI
Figure II-7. Distribution of Liberal Democrats by Country
36
Who Is More Likely to Embrace Liberal Democracy?
An attitude, such as support for the liberal exercise of democracy, cannot have just one
single factor determining it. Attitudes are influenced by personal characteristics, such as gender,
education, and place of residence, as well as formative life experiences, personality traits and
other attitudes such as, in this particular case, evaluations of the political system. It is useful to
distinguish all these different influences when trying to determine their impact on support for
liberal democracy. Thus, I develop five models that explain membership in the liberal
democracy cell as opposed to those who fall in the illiberal democracy cell or the
authoritarian cells (to simplify the statistical analysis and the discussion, liberal and illiberal
authoritarians are aggregated into a single category).
The first model includes sociodemographic characteristics: gender, age, education,

material wealth (a way to measure socioeconomic status), residence in rural or urban areas, and
whether the respondent grew up in a rural area or in a town or city. In addition to these factors, I
include two variables associated with interest in and knowledge of politics.8 This is our baseline
model.
The second model includes, in addition to the previous variables, a series of questions
associated with Putnams work on social capital (Putnam 1993) and Ingleharts (1977; 1988)
theory that general satisfaction with ones life is associated with stronger democratic attitudes.
The core of the social capital theory is that a stronger associational life and trust in others foster
the development of democratic values. Accordingly, I include a scale of participation in civil
society organizations, a question probing the respondents level of interpersonal trust, and another
inquiring about the respondents level of satisfaction with his or her life.
The third model tries to determine whether perceptions of personal or economic threat
affect our redefined regime choices. The underlying theory is that threat perception increases a
persons propensity to support authoritarian alternatives (Altemeyer 1996; Fromm 1941; Stenner
2005). The longing for security and predictability is thought to be a psychological mechanism
that leads people with an elevated threat perception to endorse authoritarianism. To measure the
sense of personal security I employ three variables. The first question asks the respondent to
assess how safe she feels in her neighborhood. The second question probes whether the
respondent has been a victim of a crime in the previous year. The last variable is an index of the
total number of corruption acts that the respondent experienced in the year previous to the survey.
To measure economic insecurity, two standard questions are utilized: the respondents
retrospective assessment of both his or her personal situation and the countrys. The assumption
is that those who have negative views of their personal or the countrys economic situation will
8
Interest in politics is measured by the answer to two questions: a) How much interest do you have in politics: a lot,
some, little or none; b) How frequently do you watch the news on TV: every day, one or twice a week, rarely, or
never? The responses were recoded so they would go from low to high interest. Political knowledge was measured
by a scale that added the correct answers to five questions of general political knowledge (see battery of GI items
in the questionnaire). Political knowledge has been found to be an important predictor of many political attitudes
(Althaus 2003; Delli Carpini and Keeter 1996; Grofman 1995).
37
have a lower sense of economic security than those who have a more positive evaluation of the
economy.
While the previous model centers on threat perceptions, the fourth model focuses its
attention on evaluations of political performance. As Bratton and Mattes (2001a; 2007) argue,
there is evidence that support for democracy is partially driven by assessments of how well the
political system is functioning. The impact of two types of political assessments, one that is
specifically related to the incumbents performance and the other that is more generally related to
the performance of the political system is explored. In the latter case, three variables are used: a
scale of system support, a scale of trust in political institutions, and a question that measures the
respondents satisfaction with the way in which democracy functions in his or her country.
Finally, the fifth model adds to all the previously mentioned influences two variables that
measure ideology (on the traditional left-right scale) and a scale of social authoritarian values
(which, as I mentioned before, it is not contaminated with political content). Following the
work of Altemeyer on conventionalism (1996), I hypothesize that those who are on the socially
authoritarian side the spectrum (displaying support for patriarchal and authoritarian conventions)
are more likely to support authoritarian forms of government.
The table reporting the full results for the five models is attached at the end of the chapter.
Table I-5 summarizes the results in a non-technical fashion, providing the sign of the relationship
for only those predictors that turn out to be statistically significant. Predictors that are not
significant are noted as ns. The reference category in the table is those who support
authoritarianism (whether in its liberal or illiberal variants).
To analyze the results, I will compare, first, the factors that differentiate those who
support liberal democracy from those who endorse authoritarian forms of government. After that,
I will analyze the differences between those who support illiberal forms of democracy versus
those who endorse authoritarianism.
The summary table suggests that three sociodemographic factors are consistent predictors
of support for liberal democracy. Older citizens are more likely to be classified as liberal
democrats (based on their answers to the questions on regime preference and how executive
power should be exercised) than younger voters (Figure II-8). This finding is entirely
understandable but nonetheless troublesome. Older citizens grew up in a period when many
Latin American countries were experiencing military rule. They probably remember well the
abuses associated with these regimes and the citizens efforts to dislodge them from power.
Younger citizens, on the other hand, were socialized in a different environment. They never
experienced regimes other than the often weak and troubled democratic governments that
characterize a good part of the region; therefore they lack the comparative reference afforded to
those who had lived under authoritarian regimes. Not surprisingly, the association between age
and regime choice is very consistent and remains significant even when variables associated with
political performance and democratic values are introduced in the analysis.
In three of the five models gender emerges as a statistical significant predictor, with men
more likely than women to endorse liberal democracy than authoritarianism. Similarly, the data
38
lend general support to the idea that greater possession of material goods enhances democratic
predispositions.
The evidence also shows that liberal democrats, in comparison to authoritarians, are more
educated and more likely to have spent their formative years in towns or cities than
authoritarians. These findings are consistent with what one would expect, as both education and
early socialization in urban areas are thought to be related with stronger democratic preferences.
39
Table II-4. Predictor Variables of Support for Liberal and Illiberal Democracy (Summary)
Baseline Social Threat Political Values
Predictor Variables Model Capital Perception Performance
I. Liberal Democrats
Age in years + + + + +
Gender (0=Women) ns + + + ns
Education in years + + + + +
Material wealth + + + + ns
Hometown (0=countryside) + + + + +
Hometown (0=city) ns ns ns ns ns
Place of residence (0=rural areas) ns ns ns ns ns
Watch news on TV ns ns ns ns ns
Interest in politics ns ns ns ns ns
Political knowledge + + + + +
Associational life ns ns ns ns
Interpersonal trust + + + ns
Life satisfaction + + ns ns
Feels safe in his/her neighborhood ns ns ns
Has been a victim of crime (0=No) ns ns ns
Index of corruption victimization - - -
Retrospective eval. of personal economic situation ns ns ns
Retrospective eval. of countrys economic situation - - -
Presidential approval ns ns
System support + ns
Trust in political institutions ns ns
Satisfaction with the way democracy works + +
Ideology ns
Scale of authoritarian values -
Scale of political tolerance +
II. Illiberal Democrats
Age in years + + + + +
Gender (0=Women) ns ns ns ns ns
Education in years ns ns ns ns ns
Material wealth ns ns ns ns ns
40
Hometown (0=countryside) ns ns ns ns ns
Hometown (0=city) ns ns ns ns ns
Place of residence (0=rural areas) ns - - - -
Watch news on TV ns ns ns ns ns
Interest in politics ns ns ns ns ns
Political knowledge + + + + +
Associational life ns ns ns ns
Interpersonal trust ns ns ns ns
Life satisfaction + + ns ns
Feels safe in his/her neighborhood ns ns ns
Has been a victim of crime (0=No) ns ns ns
Index of corruption victimization - ns ns
Retrospective eval. of personal economic situation ns ns ns
Retrospective eval. of countrys economic situation - - -
Presidential approval - -
System support ns ns
Trust in political institutions ns ns
Satisfaction with the way democracy works + +
Ideology ns
Scale of authoritarian values -
Scale of political tolerance ns
Reference category: Authoritarians; ns=not significant. Variables not included in the model are shaded.
Neither TV news watching nor interest in politics increases the likelihood of choosing
liberal forms of democracy over authoritarian governance. On the other hand, and consistent
with our previous finding about the role of education, having higher levels of political knowledge
is associated with a greater likelihood of belonging in the liberal democracy cell.
Turning to variables associated with the social capital approach, the data show very mixed
support for the view that interpersonal trust or membership in civil society organizations
strengthen commitment to democratic values. When included along with the variables that
constitute our baseline model, both interpersonal trust and associational life emerge as significant
predictors of membership in the liberal democracy group. They are also significant when they
are included in the perception of threat model. However, when political performance, ideology,
political tolerance, and social authoritarianism variables are taken into account, the effect of
social capital variables disappears. A similar pattern can be detected when the impact of life
satisfaction on support for liberal democracy is analyzed. The impact is positive and significant
when it is first introduced, but when performance and other attitudinal variables are included, its
influence vanishes.
41
40
39
Mean Age
38
37
36
Liberal Democrats Illiberal Democrats Authoritarians
Figure II-8. Regime Preference/Support for Liberal Rule by Age
There is also mixed support for the thesis that elevated levels of threat perception depress
support for liberal forms of democracy. Whether a person has been victim of a crime does not
make him or her less likely to support democracy. Nor is the sense of personal insecurity
associated with regime preferences. This is relatively good news because it implies that the
increase in crime that the region is experiencing is not affecting, at least not yet, attitudes towards
liberal democracy. On the other hand, I do find that greater personal experience with corruption
and negative evaluations of the national economy diminish the likelihood that a respondent will
prefer liberal democracy over authoritarianism. People who believe that the countrys economic
is worse today than it was in the previous are less likely to support liberal democracy and more
likely to embrace authoritarianism (Figure II-9). This effect does not disappear when political
performance and other attitudinal variables are included in the analysis.
42
70%
Better
65%
Same
Worse
60%
55%
50%
45%
%
40%
61.1% 62.2%
35% 58.0%
30%
25%
20% 28.5%
25.6%
24.1%
15%
13.3% 13.7% 13.5%
10%
Barras de error: 95% IC
Figure II-9. Regime Preference/Support for Liberal Rule by Retrospective

Evaluations of the Countrys economy
Model 4 examines how important are assessments of the political system, trust in political
institutions, presidential approval and general satisfaction with the way democracy is working in
the support for liberal democracy. The results indicate that preference for liberal democracy is
not affected by assessments of the incumbent performance or trust in political institutions.
Similarly, while system support emerges as a positive, significant predictor in one model, its
effect vanishes when ideology, political tolerance, and the scale of authoritarian attitudes are
included in the analysis. Of the political performance variables, only one exhibit a consistent
influence on regime choice, namely satisfaction with the way democracy is working: a
generalized sense that democracy is not functioning properly makes people more likely to support
authoritarianism (Figure II-10).
Finally, while we fail to find a relationship between ideology (measured on a left-right

scale) and our redefined regime categories, we do find that rejection of social authoritarian values
increase support for liberal democracy.
43
50
Mean Satisfaction with Democracy 48
46
44
42
Figure II-10. Regime Preference/Support for Liberal Rule by Satisfaction

with the way Democracy is Working
Turning our attention to the differences between liberal and illiberal democrats, the
following can be said. First of all, many of the factors that help us discriminate between liberal
democrats and authoritarians help us also discriminate between illiberal democrats and
authoritarians. As the summary table suggests, older citizens with levels of political knowledge
higher than the average, who have positive assessments of the countrys economy, generally
satisfied with the way democracy works, and who are also politically tolerant and less inclined to
support social authoritarianism are more likely to endorse democracy in illiberal variants than to
support authoritarianism.
However, some of the factors that make people choose liberal democracy as opposed to
authoritarianism do not help us distinguish between those who prefer illiberal forms of
democracy and authoritarianism. For instance, while having early socialization experiences in
towns or cities as opposed to rural areas lead people to endorse liberal democracy, these
socialization experiences do not help us distinguish between illiberal democrats and
authoritarians. In similar vein, while people with more material possessions are more likely to
choose liberal democracy as opposed to authoritarianism, economic wealth does not help us
discriminate between authoritarians and illiberal democrats. This means that authoritarians and
illiberal democrats tend to have similar early socialization and economic wealth backgrounds.
On the other hand, there are two variables that are significant predictors of membership in
the group of illiberal democrats as opposed to authoritarians. The first is the area of current
residence, with illiberal democrats being more likely to live in cities than authoritarians are. The
44
second variable is presidential approval. Those who are unsatisfied with the job of the president
are more inclined to support authoritarianism than to endorse illiberal forms of democracy.
Those who approve of the job the president or the prime minister is doing are more likely to
endorse democracy but in its illiberal variant. This makes sense because illiberal democrats are
the ones willing to grant extraordinary powers to presidents or prime ministers. Liberal
democrats, on the other hand, may or may not support the president. Whether they do or do not
has no bearing on their disposition to prefer one type of rule over other.
Concluding Remarks
I find that the majority of Latin Americans define democracy in normative terms, and that
those who do so are more likely to endorse it than those who define it in instrumental or negative
terms. But I also find that the effect is not very strong. The data show that the majority of Latin
Americans endorse democracy, but that there are important cross-national variations. Moreover,
the study finds that this support is rather shallow. Many of the self-declared democrats hold
views that are inimical to democratic rule, such as endorsement of military coups or willingness
to support illiberal exercises of political power. I find that depending on the question, as few as
17 percent and as many as 45 percent of self-declared democrats are ready to justify a military
coup. I also find that a quarter of those who endorse democracy have no problem agreeing with
the statement that presidents should limit the voice and vote of the opposition if it is necessary
for the progress of this country.
Given these inconsistencies, I proposed a typology that seeks to provide a more robust
conceptualization of support for democratic rule. I did so by combining regime preferences (for
democracy or authoritarianism) with attitudes about how political power should be exercised
(with constraints or without constraints). Liberal democrats, in this conceptualization, are those
who support democracy in principle and are unwilling to grant unchecked powers to presidents or
prime ministers. Illiberal democrats are those that while preferring democracy are also willing to
endorse the illiberal exercise of power. Liberal and illiberal authoritarians are, for purposes of
this article, grouped into a single category. While 73 percent of the respondents declare a
preference for democracy over other regimes, the analysis conducted in this chapter shows that
only 61 percent of all respondents could be classified as liberal democrats. The latter figure
probably provides a more realistic picture of the extent of attitudinal support for democracy than
the former. Even then, and as the subsequent analysis makes clear, some liberal democrats are not
completely exempted from the temptation of endorsing some authoritarian positions.
To identify who are those more likely to be classified as liberal democrats, a series of
models were developed. The main findings are that endorsement of liberal democracy is driven
by both instrumental and normative considerations. On the utilitarian side, the data show that
positive evaluations of the countrys economy and general satisfaction with the way democracy is
working increase support for liberal forms of democratic rule. On the normative side, the data
show that greater levels of political tolerance and low predispositions to social authoritarianism
increase support for liberal democracy. In addition to these utilitarian and normative
considerations, certain demographic characteristics such as age and material wealth, as well as
political information, affect the degree of support for liberal democracy, with older, more affluent
45
and more politically informed citizens being more likely to prefer liberal democracy than
authoritarianism.
The general conclusion of this chapter is that mass support for democracy is not as high or
robust as one would expect or hope. On the other hand, it is clear than a majority of Latin
Americans look with sympathy on the liberal exercise of democratic rule. Even if this support is
at times inconsistent, it provides a core from which stronger and more consistent democratic
attitudes could be encouraged.
46
References
Altemeyer, Bob. Enemies of Freedom: Understanding Right-Wing Authoritarianism. San

Francisco: Jossey-Bass, 1988.
. The Authoritarian Specter. Cambridge: Harvard University Press, 1996.
Althaus, Scott L. Collective Preferences in Democratic Politics: Opinion Surveys and the Will of
the People. Cambridge: Cambridge University Press, 2003.
Bratton, Michael and Robert Mattes. Support for Democracy in Africa: Intrinsic or
Instrumental. British Journal of Political Science. Vol. 31, No. 3 (2001a): 447-474.
. Africans Surprising Universalism. Journal of Democracy 12 (1) (2001b): 107-121.
Cameron, Maxwell. Self-Coups: Peru, Guatemala, Russia. Journal of Democracy 9 (1) (1998).
Carrin, Julio F. Conclusion: The Rise and Fall of Electoral Authoritarianism in Peru. In The
Fujimori Legacy: The Rise of Electoral Authoritarianism in Peru, ed. Julio F. Carrin,
University Park: Pennsylvania State University Press, 2006.
Conaghan, Catherine M. Fujimoris Peru: Deception in the Public Sphere. Pittsburgh: University
of Pittsburgh Press, 2005.
Dahl, Robert. A Preface to Democratic Theory. Chicago and London: The University of Chicago
Press, (1956).
. Democracy and its Critics. New Haven: Yale University Press, 1989.
Delli Carpini, Michael X. and Scott Keeter. What American Know About Politics and Why it
Matters. New Haven: Yale University Press, 1996.
Diamond, Larry. Developing Democracy: Toward Consolidation. Baltimore and London: The
Johns Hopkins University Press, 1999.
. Thinking about Hybrid Regimes. Journal of Democracy 13 (2) (2002).
Fromm, Eric. Escape From Freedom. New York: Holt, Rinehart and Winston, 1941.
Grofman, Bernard (editor). Information, Participation, and Choice. Ann Arbor: University of
Michigan Press, 1995.
Held, David. Models of Democracy. Second Edition. Stanford: Stanford University Press, (1996).
Inglehart, Ronald. The Silent Revolution. Princeton: Princeton University Press, 1977.
. The Renaissance of Political Culture. American Political Science Review. Vol. 82
(1988): 1203-1230.
Lakoff, Sanford. Democracy: History, Theory, Practice. Boulder: Westview Press, (1996).
Levitsky, Steven, and Lucan A. Way. The Rise of Competitive of Authoritarianism. Journal of
Democracy 13 (2) (2002).
Macpherson, C.B. The Life and Times of Liberal Democracy. Oxford: Oxford University Press,
1977.
Mattes, Robert and Michael Bratton. Learning about Democracy in Africa: Awareness,
Performance, and Experience. American Journal of Political Science 51 (1) (2007): 192-
217.
47
PNUD-PRODDAL. La democracia en Amrica Latina: Hacia una democracia de ciudadanos y

ciudadanas. Lima: PNUD, 2004.
Prothro, James W. and Charles M. Grigg. Fundamental Principles of Democracy: Bases of
Agreement and Disagreement. Journal of Politics 22 (2) (1960): 276-294.
Putnam, Robert. Making Democracy Work. Princeton: Princeton University Press, 1993.
Sarsfield, Rodolfo. La no-eleccin de Dorian Gray o la decisin de Ulises? Racionalidad y
determinacin en la preferencia por democracia en Amrica Latina. Mxico: FLACSO.
Ph.D Dissertation, 2003.
Sarsfield, Rodolfo and Julio F. Carrin. The Different Paths to Authoritarianism: Rationality and
Irrationality in Regime Preferences. Annual Meeting of the World Association of Public
Opinion Research, WAPOR. May 16-18, Montreal, Canada, 2006.
Sarsfield, Rodolfo and Fabian Echegaray. Opening the Black Box. How Satisfaction with
Democracy and its Perceived Efficacy Affect Regime Preference in Latin America.
International Journal of Public Opinion Research, Vol. 18 (2006): 153-173.
Schedler, Andreas (editor). Electoral Authoritarianism: The Dynamics of Unfree Competition.
Boulder and London: Lynne Rienner Pub, 2006.
Stenner, Karen. The Authoritarian Dynamic. New York: Cambridge University Press, 2005.
Touraine, Alain. What is Democracy. Boulder, Westview Press, 1997.
48
Appendix. Modeling Support for Liberal Democracy

Table II-6. Predictors of Support for Liberal Democracy
I. Liberal Democrats
-1.498 -11.875 -12.218 -11.918 -9.554
Constant
(.765)* (1.240)** (1.265)** (1.292)** (1.437)**
.014 .014 .014 .015 .016
Age in years
(.001)** (.001)** (.001)** (.001)** (.002)**
.062 .080 .076 .076 .070
Gender (0=Women)
(.034) (.036)* (.037)* (.039)* (.042)
.032 .037 .038 .039 .030
Education in years
(.005)** (.005)** (.005)** (.006)** (.006)**
.057 .026 .028 .024 .020
Material wealth
(.011)** (.011)* (.012)* (012)* (.013)
.199 .192 .178 .187 .205
Hometown (0=countryside)
(.048)** (.051)** (.052)** (.054)** (.059)**
.022 .065 .071 .101 .084
Hometown (0=city)
(.048) (.051) (.052) (.053) (.058)
.022 -.051 -.038 -.030 -.031
Place of residence (0=rural areas)
(.043) (.046) (.047) (.050) (.055)
-.019 -.009 -.014 -.012 -.005
Watch news on TV
(.019) (.021) (.021) (.022) (.024)
.012 .001 -.002 -.002 -.007
Interest in politics
(.018) (.019) (.020) (.021) (.022)
.194 .113 .112 .110 .078
Political knowledge
(.014)** (.016)** (.017)** (.017)** (.019)**
-.083 -.038 -.107 .075
Associational life
(.087) (.089) (.093) (.101)
.086 .066 .047 .031
Interpersonal trust
(.020)** (.021)* (.022)* (.024)
.070 .056 .030 .027
Life satisfaction
(.024)* (.025)* (.026) (.028)
.001 .001 .001
Feels safe in his/her neighborhood
(.001) (.001) (.001)
.057 .029 .043
Has been a victim of crime (0=No)
(.049) (.050) (.055)
-.088 -.066 -.058
Index of corruption victimization
(.024)** (.026)* (.029)*
49

Retrospective evaluation of personal -.014 .029 .028
economic situation (.030) (.031) (.033)
Retrospective evaluation of -.137 -.133 -.154
countrys economic situation (.030)** (.032)** (.035)**
.046 .035
Presidential approval
(.024) (.027)
.003 .002
System support
(.001)* (.001)
-.001 -.001
Trust in political institutions
(.001) (.001)
Satisfaction with the way .005 .005
democracy works (.001)** (.001)**
.011
Ideology
(.008)
-.010
Scale of authoritarian values
(.001)**
.009
Scale of political tolerance
(.001)**
II. Illiberal Democrats
-3.017 -12.966 -13.689 -14.099 -11.590
Constant
(1.112)* (1.605)** (1.319)** (1.667)** (1.855)**
.010 .011 .011 .012 .013
Age in years
(.002)** (.002)** (.002)** (.002)** (.002)**
-.003 .049 .057 .052 .034
Gender (0=Women)
(.049) (.052) (.053) (.055) (.060)
.004 .006 .006 .007 .000
Education in years
(.008) (.008) (.008) (.008) (.009)
.025 .004 -.006 -.008 -.005
Material wealth
(.016) (.017) (.017) (.018) (.019)
.103 .110 .101 .090 .102
Hometown (0=countryside)
(.067) (.071) (.072) (.075) (.082)
-.091 -.077 -.055 -.049 -.065
Hometown (0=city)
(.068) (.071) (.072) (.075) (.080)
-.087 -.186 -.182 -.160 -.173
Place of residence (0=rural areas)
(.062) (.066)* (.068)* (.071)* (.078)*
-.018 -.008 -.018 -.018 -.016
Watch news on TV
(.028) (.029) (.030) (.031) (.034)
.044 .040 .033 .015 .014
Interest in politics
(.026) (.027) (.028) (.029) (.032)
Political knowledge .152 .087 .087 .091 .060
50

(.020)** (.023)** (.024)** (.024)** (.027)*
-.008 .034 .037 .123
Associational life
(.123) (.126) (.132) (.142)
.046 .012 -.003 -0.012
Interpersonal trust
(.028) (.030) (.031) (.034)
.123 .084 .045 .058
Life satisfaction
(.034)** (.035)* (.037) (.040)
.002 .001 .001
Feels safe in his/her neighborhood
(.001) (.001) (.001)
.095 .057 .040
Has been a victim of crime (0=No)
(.070) (.072) (.077)
-.078 -.069 -.055
Index of corruption victimization
(.034)* (.036) (.040)
Retrospective evaluation of personal -.008 .020 .018
economic situation (.043) (.044) (.047)
Retrospective evaluation of -.254 -.213 -.215
countrys economic situation (.043)** (.045)** (.049)**
-.147 -.118
Presidential approval
(.038)** (.042)*
.000 .002
System support
(.002) (.002)
-.002 -.003
Trust in political institutions
(.002) (.002)
Satisfaction with the way .006 .005
democracy works (.001)** (.001)**
.000
Ideology
(.012)
-.008
Scale of authoritarian values
(.001)**
.002
Scale of political tolerance
(.001)
Pseudo R square .082 .083 .087 .092 .111
-2 log likelihood 36008.5 32736.6 31749.4 29721.6 25196.6
Reference category: Authoritarians
Entries are multinomial logistic regression coefficients. Corresponding standard errors are between
parenthesis.
*p .05; **p .001. Country effects are omitted to save space.
51
III. Throw them All Out? Attitudes towards Political Parties

in the Americas1
Mara Fernanda Boidi 2
Abstract
Political parties are the intermediate structures between society and government in
representative democracies. When the links between citizens and parties are weak, the
fundamental intermediary role is not in place, and as a result, society and government are not
connected. This seems to be the case in Latin America, where citizens trust political parties
very little, and where other links between citizens and parties such as voting for political
parties, identifying with a party, and participation in party activities are also generally weak
Notwithstanding this discouraging outlook, in most countries in the region there is still a
reservoir of support for political parties. Overall, the belief in the need to have parties in order
for there to be democracy surpasses (in some cases by a wide margin) the levels of trust in
political parties. Most Latin Americans believe that parties are necessary for democracy,
ranging from 33 percent in Haiti to 66 percent in Uruguay) This suggests that contempt for
parties is related to the context: it would be the way that parties currently operate, rather than
rejection of them as an institution, which distances citizens from political parties.
In 1942, referring to the American party system, Schattschneider stated that modern
democracy is unthinkable without political parties (1942:1). Contemporary democracies
continue to be unthinkable without political parties because democracies have been designed
in such a way that parties are the intermediary structure between society and government
(Sartori 1976: iv). If the links between citizens and parties are exceedingly weak (as the
available data suggest is the case in many Latin American countries) the basic intermediary
task is not being fulfilled, and as a result society and government are not connected.
One of the clearest examples, not only of the breach between citizens and political
parties, but also of the rejection of the party as an institution of political mediation is the
throw them all out protest carried out in Argentina in 2001, when people poured out into the
streets to literally oust their rulers chanting Throw them all out, not a single one stays. The
1
The Americas refers to the 22 countries included in the 2006-07 round of the LAPOP AmericasBarometer:
Bolivia, Brasil, Canada, Chile, Colombia, Costa Rica, Ecuador, the United States, El Salvador, Guatemala,
Guyana, Honduras, Haiti, Jamaica, Mexico, Nicaragua, Panama, Paraguay, Peru, Dominican Republic, Uruguay
and Venezuela. This round of interviews was made possible thanks to the support from USAID. When all the
available data allow, the analysis covers all 22 countries. The central discussion on trust in political parties is
limited to the countries of Latin America and the Caribbean.
2
This article has benefited from my conversations on the topic with Florence Faucher-King and Jonathan Hiskey.
Mitchell Seligson carefully read preliminary versions and contributed valuable comments; my thanks to all three.
The final product, of course, is my sole responsibility. I wrote this chapter during academic year 2006-2007,
while I was studying in the Ph.D program in Political Science at Vanderbilt University. I would like to thank for
the financial support received from the Department of Political Science as well as from the Center for the
Americas. Likewise, extensive thanks go to the office of the Provost at Vanderbilt University.
53
protests that culminated with then president Fernando de la Rua stepping down, was begun by
groups of unemployed workers who rejected the governments economic policies. They were
subsequently joined by middle class citizens all across the country, turning a local protest into
a national mass protest (Bonnet s.d.; Dinerstein 2003; Giarraca 2002).
Notwithstanding its origins and unique motivations, the demand to throw them all
out is an expression of citizen rejection of representative democracys central institutions,
which is very much a part of the region-wide interest in populist governments discussed in the
chapter in this volume by Mitchell Seligson. In this regard, the Argentine crisis is not
exceptional. Gonzlez (2006:8) has tallied nine severe political crises involving institutional
weakening or rupture in six other countries in Latin America to date since 2000 (Paraguay
2000, Peru 2000, Venezuela 2003, Bolivia 2003 and 2005, Ecuador 2000 and 2005, and
Nicaragua 2005). If the clamor to throw them all out is a protest against the whole system, a
demand to start over, the outrage is directed at the institutions of representative democracy,
and in this regard, the targets of the complaint are none other than the political parties as the
agents in charge of organizing, channeling, and representing interests in the public sphere.
The parties themselves recognize that they are part of the problem. In their study on
Central American political parties, Achard and Gonzlez found that 64 percent of the
politicians interviewed, and 80 percent of the non-political observers interviewed, considered
that political parties were not performing well in the region. Among the reasons that
respondents gave for this poor performance were patrimonial and corrupt practices, the
existence of oligarchic party structures, and the lack of internal democracy (Achard and
Gonzlez 2004:83). On another front, Crdova Macas (2004) pointed out that among the
causes of the crisis of political parties in Latin America are their lack of transparency and their
inability or unwillingness to carry out campaign promises and citizens demands.
The level of trust in political parties is the central theme of this chapter. Previous
research has demonstrated that a low level of trust in political parties is detrimental to the
process of democratic consolidation in several ways, as described in the main section of this
paper, Why trust in political parties matters. This chapter begins with a brief description of
citizen-parties linkages in the Americas, which is followed for the main section on trust in
political parties. The chapter concludes with a discussion of the implications that low levels of
trust in political parties have for the process of consolidating democracy in the region.
Weak Citizen-party Linkages in the Americas
The relationship among citizens and parties can be analyzed from several perspectives.
Citizens can be members of party organizations or participate in activities organized by parties,
such as public meetings, caravans or rallies. Ultimately, citizens vote for a political party, in
what constitutes for many their sole act of political participation (Campbell et al. 1960). At the
level of attitudes, aside from the aforementioned trust, individuals may or may not identify
themselves with a given political party, and they also have their own ideas regarding the role
of parties in democracy. These behaviors and attitudes, particularly when considered as a
54
whole, provide a map of the citizen-party linkages in each of the countries. This chapters aims
is to outline that map.
Table III- 1. Citizen-party linkages in the Americas

Vote in Vote in Difference Party Attend Convince Convince Work for Work for
preside legislative presidential/ identification Meetings others others the party the party
ntial election Legislative % Mean (total (only (total (only
election % vote (c) (d) responde identifiers) responden identifiers)
% (b) nts)Mean Mean ts) %
(a) (e) (f) % (h)
(g)
Mexico 71.3 58.1 13.2 49.2 0.2 0.4 0.6 7.4 9.7
Guatemala 56.5 41.1 15.4 14.7 0.1 0.4 0.7 3.8 14.2
El Salvador 67.9 65.6 2.3 31.3 0.1 0.4 0.7 8.3 15.6
Honduras 82.5 70.3 12.2 44.2 0.2 0.5 0.7 19.9 27.3
Nicaragua 61.2 46.8 14.4 49.8 0.3 0.5 0.6 10.9 16.8
Costa Rica 70.8 92.7 -21.9 36.2 0.1 0.6 0.8 11.2 19.6
Panama 79.5 51.0 28.5 20.8 0.2 0.9 1.2 12.8 26.6
Colombia 60.2 46.5 13.7 28.6 0.2 0.4 0.7 11.7 25.2
Ecuador 83.8 0.1 0.5 9.3
Bolivia 90.9 0.2 0.6
Peru 91.8 91.8 0 29.9 0.3 0.7 1.0 8.0 15.4
Paraguay 67.6 0.2 0.5 12.1
Chile 70.8 66.4 4.4 25.6 0.1 0.6 0.9 3.8 9.2
Uruguay 88.9 53.3 0.2 0.6 0.8 15.2 21.6
Brazil 84.1 34.1 0.1 0.7 0.9 14.0 23.8
Venezuela 76.8 29.8 47 32.5 0.3 0.7 1.0 13.2 26.8
Dominican 80.3 77.2 3.1 60.4 0.5 1.0 1.2 20.0 27.3
Rep.
Haiti 78.0 72.5 5.5 37.9 0.4 0.5 0.7 15.4 23.7
Jamaica 48.0 33.1 14.9 47.1 0.3 0.6 0.9 6.8 12.9
Guyana 78.1 56.8 21.3 19.5 0.3 0.8 1.1 10.1 22.5
Canada 84.0 50.7 0.2
United States 89.3 62.2 0.2
Total Americas 74.8 59.2 37.1 0.2 0.6 0.9 11.3 20.01
a. Percentage of respondents who answered yes to question vb2 Did you vote in the previous presidential election? b. Percentage of respondents
that answered yes to question vb6. Did you vote for a deputy in the last elections? c. Percentage of respondents that answered yes to question
vb10. Do you currently identify with a political party? d. Average of attendance to party meetings or political movements (0. Never, 1. Once or
twice a year, 2. Once or twice a month, 3. Once a week). Created based on recodification of the original values for question cp 13. Do you attend
meetings of a political party or political movement? e. Average of how often respondents said they have attempted to convince others to vote for a
party or candidate (0. Never, 1. Rarely, 2. Sometimes, 3. Frequently). Created based on recodification of original values of question pp 1. How
frequently have you tried to convince others to vote for a party or candidate? Calculated for the total of respondents who answered the question. f.
Average of how often respondents said that they have attempted to convince others to vote for a party or candidate (0. Never, 1. Rarely, 2.
Sometimes, 3. Frequently). Created based on modification of original values of question pp 1. How often have you tried to convince others to vote
for a party or candidate? Calculated only for those identified with a political party. g. Percentage of respondents that answered yes to question pp
2. There are people who work for some party or candidate during election campaigns. Did you work for any candidate or political party during the
last presidential elections? Calculated for the total of respondents who answered the question. h. Percentage of respondents who answered yes to
question pp2.There are people who work for some party or candidate during election campaigns. Did you work for any candidate or political party
during the last presidential elections? Calculated only for those who identify with a political party.
The study in comparative perspective of behaviors such as voting for political parties
or membership in party organizations is problematic in that variations among countries could
be determined to a great extent by characteristics that are specific to the electoral systems and
the incentive structures in each country. Actually, the proportion of citizens that become
55
involved in political parties differed greatly from country to country (Table III-1). Thus,
participation through voting for parties in presidential elections can be as low as 48 percent, as
seen in Jamaica or over 90 percent as in Bolivia and Peru.3
An even greater gap between countries is found in participation in legislative elections;

participation ranges from 33 percent in Jamaica to over 90 percent in Costa Rica and Peru.
According to Mainwaring and Scully (1995:5, 9), the difference between the proportion of
votes in presidential and legislative elections is an indicator of how much political parties have
penetrated society. When political parties are the central actors in shaping preferences in
society, the differences in voting for both types of elections should be minimal given that
citizens votes would be more frequently based on party distinctions in each instance.
According to this indicator (whose results per country are presented in the third column of
Table III-1), parties would be less entrenched in Panama, Guyana, and Costa Rica, where the
differences in the proportion of voters in one and the other elections is greater. However, it is
worth mentioning that this indicator needs to be pondered in light of each countrys electoral
regulations. Whether voting is mandatory or not, and the enforcement or not of the sanctions,
as well as the concurrence in time of presidential and legislative elections are all factors that
may also affect the differences in the proportion of voters in each kind of election.
Party identification in its classic conception, according to the Michigan model, is a

psychological attachment to a political party which can occur without a formal link such as
membership, with the party. It is an affective orientation that persists through time though not
necessarily associated with consistently voting for the party that is the object of identification
(Campbell et al. 1960: 121-123). Party identification is a central link between citizens and
parties. Low levels of party identification suggest that political parties are incapable of
establishing medium and long-term links with citizens. The weakness in these links can be due
to the existence of a fluid party system in which parties appear and disappear between
elections, the inability of old and stable parties to inspire long-term loyalty, or to new
citizen values that are not reflected by the political parties and consequently do not inspire
belonging (Dalton 2006, Dalton and Wattenberg 2000). Whatever the reason, a low level of
identification with parties is an indicator of the weakness of the links between citizens and
political parties. A little over one third (37.1 percent) of citizens of the countries studied
identify with a political party (Table III-1). In this respect, variations between countries are
also significant; the proportion of citizens that identify with a party can be as low as 15 percent
in Guatemala, or as high as 60 percentage points, as in the United States and the Dominican
Republic. Other activities through which citizens connect to parties, such as attending party
rallies or participating in campaigns (also presented in Table III-1) are even less frequent
among the citizens of the Americas, although there are per-country variations in this case as
well.
3
The figures corresponding to participation in presidential and legislative elections show higher rates of
participation than those officially registered by country. This is a frequent problem in public opinion research and
it is related to the desirability effect. Social desirability refers to the need for presenting oneself in the most
favourable light possible (Turangeau and Rasinski 2000:5). Given that the act of voting is valued positively in the
democratic context of the region, some respondents could have been tempted to answer that they had voted in the
previous elections, when in reality they had not. Nevertheless, this effect does not obnubilate the important
differences among countries.
56
Prominent among the functions that political parties are typically expected to carry out
are channelling interests and demands, and the promotion of participation in elections by
reducing the cost of information at the time of voting (Mainwaring and Scully 1995: 23).
Political parties in Latin America and the Caribbean do not seem to be performing these tasks
effectively. When respondents were asked about their most important consideration when
choosing a president based on party, the candidate or the government program, the least
mentioned reason was the candidates political party (19 percent)4, and the most frequent
reason in all countries was the government plan (54 percent in the aggregate). That the
majority of citizens cast their vote for president based on the government program does not per
se imply a weakness in the party links. However, the fact that parties are the least mentioned
reason as a determiner of voting in the region as a whole, and in 12 of the 17 countries where
information is available indicates that parties are not functioning effectively in organizing and
channelling the competition for public office. For the majority of citizens, the differences are
marked by programs and candidates to a much greater extent than by political parties. As a
whole, the citizens of Latin America and the Caribbean (no data is available for Canada and
the U.S.) do not perceive political parties as fulfilling their function, namely the aggregation of
interests and information sourcing that their role presupposes (Mainwaing and Scully 1995,
Sartori 1976).
In sum, the outlook is one of relatively weak citizen-party linkages in electoral

participation, involvement with parties and participation in party and campaign activities.
There are, of course, significant differences among countries. As the information in Table III-1
suggests, the links would be stronger in Honduras, Uruguay and the Dominican Republic and
extremely weak in Guatemala. This individual level data is confirmed by the institutional
indicators: Guatemala has one of the least institutionalized party systems in the region,
whereas Honduras, Uruguay and the Dominican Republic have party systems that are much
more consolidated and established (Payne et al. 2003; Payne et al. 2006).
Why Trust in Political Parties Matters
This section draws its title from the book by Marc J. Hetherington Why Trust Matters
(2005) which looks at citizen trust in the U.S. federal government. Hetheringtons results
suggest that trust has an impact both at the time of choosing and voting for a presidential
candidate, as well as at the moment of supporting redistributive public policies. The central
argument is that trust in the government is critical when citizens feel they are paying the costs
of governmental actions without receiving the benefits. In this sense, governments require
great reserves of confidence in order for their redistributive programs to prosper. Although
Hetheringtons themes are not directly linked to the problems considered in this chapter, the
underlying concern is shared; in order to function correctly, the author says, representative
4
All respondents who asserted they voted in the last presidential elections were asked: (VB8) When you voted,
what was the most important reason you voted for: the candidates personal qualities, the candidates political
party, the candidates government plan? This question was asked in Colombia, Peru, Chile, Costa Rica, El
Salvador, Panama, Guatemala, Mexico, Haiti, Nicaragua, Paraguay, Guyana, Dominican Republic, Uruguay,
Venezuela, Honduras and Jamaica.
57
democracies require that people trust their representatives as individuals and trust in the roles
that they occupy. When people do not trust their institutions, they have no incentives to adhere
to the rules established by such institutions (Hetherington 2005:12)
Previous studies have found that citizens with low levels of trust in political parties are
less likely to vote or to participate in campaign activities; those who do not trust parties have
fewer incentives to participate in precisely the activities that involve political parties.
Consequently, these citizens have a greater propensity for seeking access to politics by non-
partisan means, such as direct contact with politicians, or direct action (marches and protests),
or they may also do so by voting for anti-party and even anti-system options (Blackelock
2006; Dalton 2006). In extreme cases, those who least trust political parties are more likely to
support non-conventional means of political participation, including illegal forms of
participation. Indeed, mistrust of parties increases electoral volatility (Dalton and Weldon
2005), which in some cases can lead to political instability and lack of ability to govern.
According to the World Values Survey (WVS), political parties are the least trusted
institution in the world. On a scale of zero to four, where zero represents no confidence and
four represents a lot of confidence, the average value of trust in political parties is 2.08. For the
rest of the institutions the mean values are: unions 2.25, parliaments 2.32, press 2.39,
government 2.4, television 2.5, the justice system 2.51, the police 2.56 and the church 2.91
(Figure III-1). The data correspond to the pooled dataset of the four waves (World Values
Survey 2005).
58
3.00
Mean (0 No trust at all- 4 Lots of trust)
2.50
2.00
1.50
2.81
2.50 2.51 2.56
2.32 2.39 2.40
2.25
1.00 2.08
0.50
0.00
Political parties Congress Government Justice System Church
Unions Press TV Police
Error bars : 95% CI World Values Survey 2005
(pooled dataset)
Figure III-1. Trust in political institutions worldwide
Cultural modernization theory explains the weakening of the ties among citizens and
parties in the industrialized world as a consequence of general changes in citizens preferences.
It assumes that high levels of economic development need to be achieved for the attitudes that
are least favorable towards parties to occur (Dalton 2006, Dalton and Wattenberg 2000, Dalton
and Weldon 2005, Inglehart and Welzel 2005). The Latin American reality, however,
undermines this theory at its most basic level: in Latin America and the Caribbean, where high
levels of economic development are yet to be experienced, the weakest links between citizens
and parties regard the issue of trust (Figure III-2). It becomes evident, therefore, that the
approach developed for the industrialized democracies is not a global explanation for the
weakness of the links between citizens and parties. What, then, explains such low levels of
trust in political parties? What other expressions can be used to evaluate the strength of the
links between citizens and parties?
59
Mean (0 No trust at all - 4 Lots of trust) 2.50
2.00
1.50
2.24 2.30
1.00 2.08
1.90 1.95
0.50
0.00
Latin America Europe North America Africa Asia

and the
Caribbean
Region
World Values Survey 2005
(pooled dataset)
Error bars: 95% CI
Figure III- 2. Trust in political parties worldwide
Levels of trust in political institutions are generally low throughout the region,
particularly when compared to trust in non-political institutions such as the Church, the media
and the military (Figure III-3). In each of the countries studied, the respondents were asked to
what extent they trusted each of these and other institutions. They were asked to indicate the
number that best described their position on a one to seven scale in which one meant no trust at
all and seven meant a lot of trust. In order to facilitate the analysis, the responses to these
questions were recoded into a zero-100 scale in which the lowest value was zero, meaning no
trust at all, and the highest, 100, meant a great deal of trust. Values over 50 indicate relative
trust in the institution, as they manifest, on average, confidence levels over the halfway point
on the scale.
Figure III-3 presents the average values of trust in each of the eight institutions in all of
the countries in the region. None of the political institutions has an average confidence level of
over the mid-point on the scale (indicated by the horizontal line at the 50 confidence level) and
political parties rated the worst. Not only are political parties the least trusted institution
among Latin America and the Caribbean citizens as a whole but also they are the institution
that inspires the least degree of confidence in each one of the countries (except in Mexico and
Venezuela where parties are the second least trusted institution following the police).
60
The second institution that inspires the least degree of confidence, the Congress, is
intimately associated with political parties, as it is their field of action. In fact, the correlation
between the levels of trust and the two institutions is relatively high (Pearsons correlation
coefficient of .508, for the entire sample), which suggests that the levels of trust in one and the
other are related.
60.00
Trust in Institutions
40.00
67.8
60.1 61.8
46.2 49.2 49.3

20.00
44.8 44.9
34.8
0.00
Court
Armed
Catholic
Political
Media
Church
Police
Parties
Congress
Forces
Elections
Supreme
Government
AmericasBarometer 2006,
Error bars: 95% CI by LAPOP
Figure III-3. Trust in institutions in Latin America and the Caribbean

To what extent do you trust? (0 Not at all, 100 A lot) Mean, by institution5
5
The values for Figure III-3 represent the average trust in each institution in the region (cases are weighed by
country, so that all countries weigh the same in the sample and therefore they weight the same in any percentage
analysis). It is noteworthy that not all questions regarding trust in the institutions were formulated in all countries;
therefore some of the values presented herein are not representative of the totality of countries included in the
2006 round of the AmericasBarometer. Specifically, trust in elections excludes Mexico, Bolivia, Chile,
Dominican Republic and Haiti; trust in the Armed Forces excludes countries where there are no standing armies,
that is Costa Rica, Panama and Haiti; finally trust in the government excludes Bolivia.
61
Guyana 48.5
Uruguay 43.3
Mexico 43.1
Guatemala 40.0
Chile 39.6
Panama 39.6
Colombia 37.9
Jamaica 36.1
Country
Costa Rica 35.9

Dominican Republic 35.2
El Salvador 35.1
Honduras 33.3
Nicaragua 32.8
Peru 32.3
Venezuela 31.5
Bolivia 31.1
Brazil 30.3
Haiti 30.0
Paraguay 25.5
Ecuador 15.1
0.00 10.00 20.00 30.00 40.00 50.00
Mean To What extent do you trust political

parties?
AmericasBarometer,
by LAPOP
Error bars: 95% CI
Figure III-4. Trust in political parties by country
What explains the low levels of trust in political parties in the region? There are at least
two clear lines of interpretation. The first one points directly to poor party performance. This
explanation emerges even in inner party circles (Achard and Gonzlez 2004) and it is also
supported by academic analyses: parties do not function well internally nor in relation to the
rest of the relevant actors in society (Crdova Macas 2004). On the other hand, there are
unmet citizen expectations. Citizens channel their frustrations arising from unmet expectations
to political parties, the visible face of democratic government. In a concept called the Central
American paradox, which can certainly be extended to the region as a whole, Achard and
Gonzlez (2004:127) point out that although there is currently more democracy than before,
citizens are less satisfied with its results, the institutions in general, and political parties in
particular. According to this thesis, citizens expectations have increasedin certain cases fed
by the unrealistic promises of political partiesbeyond the actual possibilities the systems
might have of satisfying them. This has lead to increasing disillusionment in the system in
62
general and political parties in particular. In the same vein, Paramio (1999:12) holds that
frustration of voter expectations appears as a decisive factor when explaining citizens low
opinion of party politics.
What are the most important factors affecting trust in political parties? Is it possible to
find patterns in the region in spite of the differences among countries? In the classic literature
on trust in government, the perception of institutional efficacy plays an important role: citizens
perceive that the better their leaders perform the more they can trust them, and perceptions of
economic progress produce the same reaction (Hetherington 1998; Hetherington 2005).
Consequently, it becomes imperative to analyze the impact of citizen opinions in both
dimensions government efficacy and evaluation of the economy when estimating the factors
that determine trust in parties.
Prior studies that included questions about government efficacy investigated citizens
opinions regarding the efficacy of the government in areas that interviewees themselves had
previously defined as prioritary (Hetherington 1998; Muller et al. 1991). The 2006-2007 round
of the AmericasBarometer formulated a series of questions regarding perceptions of
government efficacy, directly asking interviewees (independent from their prior answers
regarding the problems they considered most important) to what extent they believed the
current government fights poverty, promotes and protects democratic principles, battles
corruption within the government, safeguards human rights, improves citizen safety, and
combats unemployment. For each of these activities the respondents were asked to indicate
their opinion on a scale from one to seven, in which one meant nothing and seven meant a lot.
Following the standard LAPOP practice of unifying scales, the answers were re-scaled to a
continuum from zero to 100 to facilitate the analysis.
In order to measure the effect of the evaluation of governments economic performance

on the level of trust in parties, two alternative measures were utilized: evaluation of the
countrys economic situation, and the evaluation of the economic situation of the familiy
(Kinder and Kiewiet 1981). It is to be expected that in both cases the more favorable
perceptions regarding the economy will lead to greater levels of trust. 6
Citizens connections to political parties should also have a positive impact on the level
of trust in them. It is particularly expected that identification with a political party should
positively affect the level trust in all parties, given that if there is one affective orientation
towards one of the parties with a psychological identification with it (Campbell et al.
1960:121) there should be a subjacent link of trust. Consequently, it is to be expected that
identification with any political party (operationalized in the affirmative answer to the question
At this time, do you currently identify with a political party?) should be associated with
greater levels of trust.
6
Interviewees were asked whether they would assessboth the economic situation in the country and their own
household economic situationsas very good, good, neither good or bad, bad or very bad. In order to facilitate
the data analysis, answers have been coded so that the higher values correspond to more favourable perceptions.
Thus, values for both variables range from one to five (one reflects a very bad score, and five a very good
one).
63
Findings from previous research suggest that trust in political institutions is related to
interpersonal trust (Inglehart 1990, Inglehart and Welzel 2005; Putnam 1993); the links
developed due to the level of interpersonal trust in the community would influence in a
positive manner the level of trust in the system as a whole. Consequently, it is relevant to
incorporate the effects of interpersonal trust on the level of trust in political parties. The 2006-
2007 edition of the AmericasBarometer asked interviewees: Speaking of people from here,
would you say that people in this community are trustworthy, somewhat trustworthy, a little
trustworthy, or not trustworthy at all? The answers were ordered on a scale from one (not at
all trustworthy) to four (very trustworthy).
In the same way, favorable attitudes towards the political system in general should be
associated to favorable attitudes towards institutions as well (Easton 1965; Mueller et al. 1982;
Seligson 2002a), which means there should also be a positive relationship between levels of
support for the political system and trust in parties. The variable that collects information on
support for the political system To what extent do you believe that the countrys political
system should be supported? is coded on a scale from zero (not at all) to 100 (a lot).
According to some scholars, anti-party sentiment can be at least partially explained by

the rejection of corrupt practices associated with political parties (Achard and Gonzlez 2004;
Crdova Macas 2004; Dinerstein 2003). In this sense, it can be expected that there will be
lower levels of trust in parties among those who perceive them as being corrupt. Unfortunately
data on perception of corruption associated with political parties in particular is unavailable.
However, there is a direct measure of individual experience with corruption. Seligson (2002b)
found that direct personal experience with corruption has a negative impact on support for the
political system, and the same type of negative relation is to be expected regarding trust in
political parties. In the 2006-2007 round of the AmericasBarometer, citizens were asked about
their experiences of being victims of attempts of bribery by the police, a public clerk, at the
city hall, at work, at the courts, in the public health clinic, and at school. In order to compile
the responses to these questions a corruption victimization index was created. This index
ranges from zero to five (the initial distribution from zero to seven is re-scaled to facilitate the
analysis).
The previously discussed throw them all out protests calling for the withdrawal of
Argentinean politicians was later echoed in Bolivia and Ecuador and became a non-
conventional way to claim to remove public officials from their positions. As a result, it is
worth asking whether participation in public protests is related to the levels of trust in political
parties. In light of the throw them all out experience, less trust in political parties should be
expected among those who have a greater propensity to participate in protests. The 2006-2007
round of the AmericasBarometer consulted interviewees about their experience in participation
in public protests; the responses were ordered on a scale of one (never participated in a protest)
to three (participated sometimes).
Finally, it is necessary to control the impact that the individuals demographic and
socioeconomic characteristics (such as sex, age, education and wealth) can have on attitudes
towards political parties. It is also pertinent to control the impact of the level of political
knowledge and the ideological self-identification (on a scale from left to right). Finally, the
64
available evidence from the North-American context suggests that exposure to mass media,
particularly TV news, negatively affects levels of trust (Hetherington 1998). Consequently,
variables on the reception of news on radio, television, papers and on the Internet are included
in order to control for this factor. With the aim of evaluating whether there are significant
differences among countries, dummy variables are included for each country except for
Mexico, which is taken as the reference category.
When dealing with the region as a whole, several factors are found to affect trust in
parties (Table III-2). According to the results from the linear regression on trust, government
efficacy in combating poverty, the fight against corruption, unemployment reduction, the
promotion of democracy, and safety, identifying oneself with a political party, interpersonal
trust, support for the political system, and experience of victimization by corruption are all
statistically significant variables. Of the control variables, only ideology, wealth, and exposure
to news on the radio have a statistically significant impact on trust in parties. The differences
among countries in relation to Mexico, the reference category, are both statistically and
substantively significant in most cases.
As expected, general support for the system and interpersonal trust are positively
related to trust in political parties. With a 0.210 coefficient, a radical change in system support
(zero to 100 points) will produce an increase of 21 points on the scale for confidence in parties.
This is the greatest possible effect that could be found. Certainly, at this point it is valid to ask
about the causal direction of the relation; is support for parties motivated by support for the
system, or in reality is support for the system a product of trust in political parties? It is
difficult to unravel this matter, and in any case there are theoretical justifications for expecting
causal relations in both directions (which likely reinforce each other). In any case, the
unquestionable result is that in the region trust in political parties and support for the countrys
political system are parallel and move in the same direction. As Mainwaring and Scully
(1995:24) point out, in order for citizens to confer legitimacy to their government they must
believe that said governmentthrough the parties at least tries to represent society.
Also in line with expectations, identification with a political party is associated with
greater levels of trust in parties as institutions. Nevertheless, the impact is not dramatic: trust
levels among party identifiers are, on the average, 5.8 points greater than among those who are
not.
With respect to those variables that have a negative effect on trust, the individuals
experiences with corruption have the most significant impact: (-1.903), which implies that, as
expected, the greater the exposure to corruption, the lower the level of trust in parties. The
impact of experience with corruption can be a reduction of up to ten points on the level of trust
in parties.
65
Table III-2. Linear Regression on confidence in political parties in Latin America and the Caribbean (*)
B Standard
error
Evaluation of economic situation of the country .425 .268
Evaluation economic situation of the family .046 .293
Efficacy of government combating poverty .052***.011
Efficacy of government in promoting democracy .036** .012
Efficacy of government combating corruption .080***.011
Efficacy of government protecting human rights -.003 .011
Efficacy of government improving security .093***.011
Efficacy of government combating unemployment .073***.011
Identifies with a political party 5.846*** .453
Interpersonal trust 1.582*** .235
System support .210***.008
Corruption victimization -1.903*** .284
Participation in public protests .328 .313
Political Knowledge -4.5E-005 .002
Ideological self-placement .247** .084
Sex (female) .577 .428
Age .014 .015
Education -.055 .062
Wealth (possession of capital goods) -.436***.131
News on the radio .551**
.195
News on the TV -.62 .250
News in the papers .148 .227
News on Internet -.15 .326
Guatemala 4.017***1.241
El Salvador -5.325***1.131
Honduras -5.542***1.166
Nicaragua -2.846* 1.179
Costa Rica -7.604***1.144
Panama 3.160**
1.198
Peru -3.391**
1.121
Chile -3.187**
1.112
Dominican Republic -13.336***1.126
Haiti -2.684 1.396
Jamaica -1.735 1.304
Guyana 9.997***1.271
Uruguay -10.422***1.188
Venezuela -7.039***1.141
Constant 6.537***1.781
Adjustment .230
ANOVA (sig) .000
Significance p < .05 (*), p < .01 (**), p < .001 (***)
(*) Cases corresponding to Colombia, Ecuador, Bolivia , Brasil and Paraguay have been excluded due
to lack of information (some of the questions that involve independent variables were not formulated
in these countries).
66
Contrary to expectations, the evaluation of the government regarding the performance

of the economy (both at the country level and the family level) had no significant impact on
trust in parties, and the perception of the efficacy of the government although statistically
significant, did not substantially alter the levels of trust in parties. Also contrary to what was
expected, participation in public protest had no effect on trust in political parties, which would
suggest that the throw them all out outcry embodies a special kind of attitude towards
parties: not all those who participate in protests in the region hold the same vision of parties, or
at least, participation in these events does not imply significantly lower levels of trust in
political parties. Neither did exposure to the media (except for the news on the radio, although
with a substantially lower impact), nor political knowledge have any relation, according to this
model, with trust in parties. Finally, none of the sociodemographic characteristics of the
respondents (excepting wealth measured in number of capital goods, although with a
substantially marginal impact) significantly affect trust in political parties.
Differences among countries (in relation to Mexico, the reference category) are
statistically and substantially significant: the level of confidence in political parties varies
according to the country of origin (again, in relation to Mexico) when the attitudinal factors
and sociodemographic controls are kept constant.
The relatively low predictive capacity of the model presented here (Adjusted R square
of 0.230) and the variations per country suggest that perhaps there are deep --even
idiosyncratic--factors that contribute to explaining the trust in parties in greater detail. The
difference in the levels of trust in political parties among countries did not merit a multi-level
analysis, much less when it was not clear whether a single country-level variable might explain
the observed variation. However, for future research it seems pertinent to consider the need to
explore the influence of institutional frameworks (formal ones such as electoral rules, and
informal ones such as the dynamics of inter-party competition) in determining citizen attitudes,
and particularly trust in parties.
Political Parties and Democratic Consolidation in the Region
What are the implications of the preceding discussion for the democracies of the
region? According to citizens views, the political parties of the region are not performing
adequately. As discussed in section two, even when the weakening of citizen-party linkages is
a world-wide trend, it is particularly severe in Latin American, where democracies are still
under the process of consolidation.
Political parties are not accomplishing their role of providing information shortcuts
for voting choices; only one fifth of the inhabitants of the region chose a presidential candidate
for partisan reasons. In fact, the proportion of citizens that allege partisan reasons as
determiner of the vote can be as low as 4 percent, as in Venezuela. Political parties are failing
to provide meaningful alternatives for leading the government; as a consequence, they do not
fulfill their representational task. For representation to take place appropriately, the
institutionalization of the party systems is a necessary condition. Without this
67
institutionalization, as Paramio (1999: 14) points out, there is no incentive structure -for the
parties nor for the citizens- for democratic representation.
An institutionalized party system implies stability in inter-party competition, existence

of parties with roots in society, acceptance of parties and elections as the legitimate institutions
through which to gain political power, and the existence of partisan organizations with at least
some degree of formal structure (Mainwaring and Scully 1995:1). When the party systems are
little institutionalized there is more room for the emergence of populist and personalistic
alternatives. Given that politicians depend on the masses to be re-elected, once in power they
may be tempted to implement the most popular short-term policies, as opposed to mid-term
policies, potentially more favorable for the country (Mainwaring and Scully 1995:22).
Trust in political parties is a key aspect when considering the prospectus for
institutionalization of the party systems. If parties are not depositaries of citizens trust, it is
difficult for them to develop stable links or to become accepted as legitimate institutions in the
competition for public office. If, as Dalton and Weldon (2005) observed, mistrust in parties
leads to electoral volatility, the party systems of the region will not become more stable unless
citizens trust them. Therefore, trust in political parties is not just a dimension of
institutionalization; it is a condition to achieve it.
What then, are the prospects for the party systems of the region given the low level of
trust in them? Notwithstanding the discouraging outlook, in most of the countries in the region
there is still a reservoir of support for parties. The 2006-2007 round of the AmericasBarometer
asked citizens about their opinions regarding the centrality of parties in democracies.
Interviewees were asked to indicate to what extent they agreed with the statement There can
be democracy without political parties. The answers to this question (originally expressed in a
1-7 scale in which one meant completely disagree and seven meant completely agree with
the statement) were inverted, so higher values show more favorable attitudes towards parties.
Following the LAPOP standard procedure, the original variable now inverted- was re-scaled
to a range from zero to 100. This new variable represents the belief in the need of parties for
democracy; the lower values (closer to zero) suggest attitudes less favorable to parties (higher
levels of agreement with the statement that democracy can exist without them), while higher
values (closer to 100) reflect more positive attitudes towards parties (rejection to the
possibility of a democracy without political parties).7
7
This modification from the original measurement of the variable does not alter, at all, the substantive
distribution of preferences. Its only purpose is to ease the analysis and the comparability with other variables that
use the same metric.
68
70.0 64.6
AmericasBarometer 2006
by LAPOP 60.5 59.5
61.4 58.6
60.0
55.0
53.2 56.6
51.2 59.4
49.7 49.7
47.6
50.0 50.3 49.8
40.4 45.3
43.3 52.0
48.5 43.1 39.6 39.6
36.1 41.2
40.0 40.0
32.3 35.9 33.3
37.9 35.1 34.8
32.5 35.2 31.5
30.0 30.0
32.8
30.3 25.5
20.0
15.1
10.0
0.0
ca
r
a
a
na
a
s
co
)
y
u
.
or
a
l
le
i
a
ep
an
o
zi
it
gu
am
ua
al
ua
el
a
er
bi
ic
ad
hi
ad
ai
a
ya
ur
xi
ra
zu
em
e
om
R
ra
H
ag
g
C
m
an
e
nd
bb
cu
B
lv
u
ru
an
e
a
M
a
at
Ja
ar
ol
en
t
P
ic
i
E
o
U
ar
os
ic
lS
P
u
H
C
V
G
in
C
C
E
om
d
an
D
A
(L
C ount ry
al
Need of parties for democracy Trust in parties
ot
T
Figure III-5. Need of parties for democracy and trust in political parties
The gray line in Figure III-5 shows the values of this variable for each country side by
side with averages of trust in parties (black line). Both trust in parties as well as need for
parties in democracies, are presented in Figure III-5 on a scale from zero to 100, and for both
variables the highest values reflect the most positive attitudes towards parties (in one case it is
trust, in the other rejection of the possibility of the existence of democracy without parties).
The values reported are country averages (for all the countries where the two questions were
formulated) and for the region as a whole. The use of the same scale for both variables allows
for direct comparison and facilitates the presentation of the differences between them,
represented by the breach between the values in both lines.
In the region as a whole, and in the majority of the countries, the belief that parties are
necessary for democracy is significantly greater than the levels of trust in parties. On average,
there are more citizens who support with greater conviction the necessity of parties for
democracy than those who claim to trust them. The difference between trust in parties and the
belief that parties are central to democracy suggests that the disdain for parties is
circumstantial. It is parties as they currently function rather than a rejection of them as an
institution that keeps citizens away from parties.
This gap between trust in parties and the relatively higher rejection of democracy
without parties seems to reflect an imbalance in the supply and demand of the democratic
institutions (Bratton 2005: 28). Bratton points out that in order for democracy to be insured, it
is necessary that popular demands be met by the supply of democratic institutions, principally
by the elite; democratic consolidation would be reached when the equilibrium is produced at a
high level of supply and demand. In the case of political parties, the most favorable situation
69
would be an increase in trust that should reach the level of expectations of party
indispensability for democracy.
Rejection of democracy with no political parties, on the other hand, is not

overwhelming. The average in the region is 52 points, the middle of the scale. When
considering individual countries, it varies from 33 points in Haiti to 65 points in Uruguay.
However, in the majority of cases rejection of democracy without parties surpasses the levels
of trust, which would indicate that parties have the opportunity to recover lost ground if they
improve their image and performance. Ideally, actions that aim at elevating public confidence
in parties would lead to a more favorable vision of parties as necessary actors in democracy, in
a dynamic process in which specific party support could feed diffuse institutional support in a
representative democracy (Easton 1965), thereby strengthening the process of democratic
consolidation in the region.
Democracies in the region need parties (as do all contemporary democracies) to

function adequately, as they were designed to operate under representative principles, by way
of channels that are institutionalized for this end. In the majority of Latin American and
Caribbean countries, the agents of this representation, namely political parties, are not
performing in accordance with citizen expectations. Nevertheless, the evidence suggests that
should parties improve their performance (fundamentally gaining citizens trust) there would
be subsequent possibilities for party representation to become strengthened thus contributing
to the strengthening of the democracies in the region.
70
References
Achard, Diego; Luis E. Gonzlez. "Bringing All Voices Together: The State and Prospects for
Political Parties Central America, Panama and the Dominican Republic." In A
Challenge for Democracy. Political Parties in Central America, Panama and the
Dominican Republic, edited by Diego Achard; Luis E. Gonzlez. Washington, D.C.:
Inter-American Development Bank, International IDEA, Organization of American
States, United Nations Development Program, 2004.
Blakelock, Paul. "Changing Trust: Individual-Level Assessments of Political Legitimacy."
Ph.D., University of Houston, 2006.
Bonnet, Alberto R. Que Se Vayan Todos. Crisis, Insurreccin Y Cada De La Convertibilidad
[cited Enero 2007]. Available from
http://www.geocities.com/economistas_de_izquierda/albertobonnet_quesevayantodos.
PDF.
Bratton, Michael; Robert Mattes; E. Gyimah-Boadi. Public Opinion, Democracy and Market
Reform in Africa: Cambridge University Press, 2005.
Campbell, Angus; Philip E. Converse; Warren Miller; Donald Stokes. The American Voter.
Chicago: The University of Chicago Press, 1960.
Crdova Macas, Ricardo. "The Crisis of Political Parties in Latin America." In A Challenge
for Democracy. Political Parties in Central America, Panama and the Dominican
Republic, edited by Diego Achard; Luis E. Gonzlez. Washington, D.C.: Inter-
American Development Bank, International IDEA, Organization of American States,
United Nations Development Program, 2004.
Dalton, Russell J. Citizens Politics. Public Opinion and Political Parties in Advanced
Industrial Democracies. Fourth ed. Washington, D.C: CQ Press, 2006.
Dalton, Russell J.; Martin P. Wattenberg . "Unthinkable Democracy. Political Change in
Advanced Industrial Democracies." In Parties without Partisans. Political Change in
Advanced Industrial Democracies
edited by Russell J Dalton; Martin P. Wattenberg. Oxford and New York: Oxford University
Press, 2000.
Dalton, Russell J.; Steven A. Weldon. "Public Images of Political Parties: A Necessary Evil?"
West European Politics 28, no. 5 (2005): 931-51.
Dinerstein, Ana C. "Que Se Vayan Todos! Popular Insurrection and the Asambleas Barriales
in Argentina." Bulletin of Latin American Research 22, no. 2 (2003): 187-200.
Easton, David. A Framework for Political Analysis. New Jersey: Prentice-Hall, 1965.
Giarraca, Norma. "Argentina 1991-2002: Una Dcada De Protesta Que Finaliza En Un
Comienzo. La Mirada Desde El Interior Del Pas." Argumentos 1, no. 1 (2002): 1-8.
Gonzlez, Luis E. "Las Crisis Polticas De Amrica Latina En Los Primeros Aos Del Siglo."
In Poltica Y Desarrollo En Honduras, 2006-2009, edited by Diego Achard; Luis E.
Gonzlez. Tegucigalpa: UNDP ASDI AECI DFID, 2006.
Hetherington, Marc J. "The Political Relevance of Political Trust." American Political Science
Review 92, no. 4 (1998): 791-808.
. Why Trust Matters. Declining Political Trust and the Demise of American Liberalism.
Princeton and Oxford: Princeton University Press, 2005.
71
Inglehart, Ronald. Culture Shift in Advanced Industrial Societies. . New Haven: Princeton
Inglehart, Ronald; Christian Welzel. Modernization, Cultural Change and Democracy:
Kinder, Donald R.; Roderick Kiewiet. "Sociotropic Politics: The American Case." British
Journal of Political Science 11, no. 2 (1981): 129-61.
Mainwaring, Scott; Timothy R. Scully. "Introduction. Party Systems in Latin America." In
Building Democratic Institutions. Party Systems in Latin America, edited by
Mainwaring Scott and Timothy R. Scully, 1-34. Stanford: Stanford University Press,
1995.
Muller, Edward N.; Henry A. Dietz; Steven E. Finkel. "Discontent and Expected Utility of
Rebellion: The Case of Peru." The American Political Science Review 85, no. 4 (1991):
1261-82.
Muller, Edward N.; Thomas O. Jukam; Mitchell Seligson. "Diffuse Political Support and
Antisystem Political Behavior: A Comparative Analysis." American Journal of
Political Science 26, no. 2 (1982): 240-64.
Paramio, Ludolfo. La Democracia Tras Las Reformas Econmicas En Amrica Latina
Instituto de Estudios Sociales Avanzados (CSIC), 1999 [cited 2007]. Available from
http://www.iesam.csic.es/doctrab1/dt-9903.htm.
Payne, J. Mark et al. Democracies in Development. Politics and Reform in Latin America.
Washington, D.C.: Inter-American Development Bank, Institute for Democracy and
Electoral Assistance, 2003.
Payne, Mark J.; Daniel Zovatto; Mercedes Mateo Daz. La Poltica Importa. Democracia Y
Desarrollo En Amrica Latina. Segunda Edicin ed. Washington, DC: IADB - IDEA,
2006.
Putnam, Robert D. Making Democracy Work: Civic Traditions in Modern Italy. Princeton:
Princeton University Press, 1993.
Sartori, Giovanni. Parties and Party Systems. A Framework for Analysis. London New York
Melbourne: Cambridge University Press, 1976.
Schattschneider, E.E. Party Government. Edited by Phillips Bradley, American Government in
Action. New York: Farrar and Rinehart, 1942.
Seligson, Mitchell A. "The Impact of Corruption on Regime Legitimacy: A Comparative
Study of Four Latin American Countries." The Journal of Politics 64, no. 2 (2002a):
408-33.
. "Trouble in Paradise? The Erosion of System Support in Costa Rica, 1978-1999."
Latin American Research Review 37, no. 1 (2002b): 160-85.
Survey, World Values. World Values Survey, Four-Waves Pooled Dataset 2005 [cited August
2006]. Available from www.worldvaluessurvey.org.
72
IV. Economic Performance and Support for the System:

Economic Challenges for Latin American Democracies
Vivian Schwarz-Blum
Abstract
The goal of this paper is to determine the effect that individual and system-lvell evaluations of
government economic performance in Latin American and Caribbean countries have on
system support.. The study makes use of a linear regression analysis in order to assess the
effects of individual evaluations, and a mixed multi-level analysis to estimate the effect of
national-level variables on support for the system. Results show that, in contrast to evidence
gathered from highly industrialized nations, in the Latin America region individual
evaluations the governments economic performance and system level factors are both
important predictors of support for the system.
The study of stability of democracies has generated a substantial amount of theory and
empirical evidence whose center of attention is the importance of support for the political
system for the stability and even the survival of democratic regimes. The dynamics of support
for the system depend, in part, on the relations between the State and the citizens of a country.
The fundamental assumption in this relationship is that citizens have expectations regarding
the States performance and that said performance is judged by citizens individually. Positive
or negative evaluations result in feelings of satisfaction or dissatisfaction with the States
performance, which in turn ought to lead to the willingnessor lack thereofin the individual
to support and have confidence in the State, and to consequently act in a cooperative manner
with the institutions and regulations generated by the State. On the other hand, unsatisfactory
evaluations ought to produce a low level of cooperation, and questioning of regulations and
decisions emanating from the State, which would weaken and even destabilize the State if the
sentiment of dissatisfaction is generalized among citizens.
Based on this reasoning, scholars have found that support for the political system is an
important element in a democracy not only owing to its relevance for the stability of the
democratic system but also because it is a vital element for the political process, given that a
democratic political system cannot survive for long without the support of a majority of its
citizens (Miller 1974).
Low levels of support for the system have been considered a threatening factor for the
Vivian Schwarz-Blum is a PhD student in the Political Science Department at Vanderbilt University and is also
part of the LAPOP crew.
A special thanks for professors Mitchell A. Seligson and Jonathan Hiskey for their valuable advice for this
study. I also thank my co-workers at LAPOP and classmates in the Political Science Department for their
suggestions and comments.
73
stability of democracies because the consequences of long-term dissatisfaction with the

government can generate, as Miller suggests, the feeling of a power vacuum and of the
absence of regulations. Feelings of powerlessness and normlessness are very likely to be
accompanied by hostility toward political and social leaders, the institutions of government
and the regime as a whole (Miller 1974).
The effect of government performance on citizens can have more serious consequences
than what is often believed. Bastian and Luckham (2003) propose that it is necessary to carry
out an in depth study of the effects of democracies on the lives of citizens because, contrary to
conventional notions, democracy is 'Janus-faced,' namely, it can generate both negative and
positive consequences, depending on the power and the performance of the political players,
the interest groups and the economic conditions. On the basis of the aforementioned factors,
democracies can emphasize social inequalities, marginalize minorities, or intensify conflicts
among social groups in the same way that they can increase the participation and inclusion of
marginalized groups in the political processes, create policies of social inclusion and improve
the redistribution of resources among the population.
Dissatisfaction with government performance can be changed into action at the

individual level through system support. Political confidence plays a key role in the perception
of legitimacy of the norms issued by a government and therefore in the degree to which an
individual is willing to support the government and the regime.
Levi and Stoker (2000) emphasize the importance of the relation between political trust
and that which is trustworthy, and the political consequences of the presence or absence of
those elements in political regimes. Their study concentrates on phenomena which reflect
confidence in the regime, such as public opinion and participation in politics and their relation
to trust, social trust, and citizen cooperation with and obedience to the incumbent government
with regard to the levels of trust it inspires.
This study concludes that trust is important for strengthening of the legitimacy of
political regimes. The fact that citizens consider the government or politicians trustworthy has
an influence on, among other things, the levels of individual participation in politics and the
probability that individuals will become politically active, as well as their electoral
preferences, the levels of social cooperation and individual support for government policies
and for the political regime.
Aside from the Levi and Stoker study, there are several studies (Easton 1975; Easton
1976; Weatherford 1992; Hetherington 1998; Schwarz-Blum 2006) that have provided ample
evidence that political trust is a fundamental part of support for the system. The LAPOP
instrument for measuring the level of support for the system of a specific country is based on
an index of five items that have been studied and established as valid by scholars and
researchers of democracy and is intended to show the level of confidence with which
interviewees consider the central elements of their political system.
These items are measured on a seven-point scale. In order to facilitate the

comprehension of the analysis, the scale has been transformed into a scale from zero to 100, in
74
which an average closer to zero is an indicator of low levels of support for the system and an
average closer to 100 is an indicator of a high level of support for the system.
The items that make up this index are measured through the following questions:
Index of Support for the System

B1. To what degree do you believe the courts in (country) guarantee a fair trial?
B2. To what degree do you respect the political institutions in (country)?
B3. To what degree do you believe that the citizens basic rights are safeguarded by the political system in
(country)?
B4. To what degree do you feel proud of living in the political system in (country)?
B6. To what degree do you think the political system in (country) should be supported?
What is the relationship between these theories and the present circumstances in the
democracies of Latin America? Many Latin American democracies belong to the group of the
so called third wave of democratization (Huntington 1991), which means that they are young
democracies that are still undergoing processes of transformation and consolidation. Many of
these young democracies were thought to have been consolidated to the point that they would
not revert to authoritarian regimes. However, the resurgence and spread of populism that
favors extremely personalized leadership in South America and Central America, indicates a
growing acceptance of socialist ideologies that reject liberal democracy as an ideal form of
government, as evidenced in the election of popular leaders such as Ortega in Nicaragua,
Morales in Bolivia and Correa in Ecuador as well as the growing popularity of the governing
style of Hugo Chavez. All of these factors concern citizens and scholars who are committed to
pure democratic principles and are once again questioning the destiny of the younger
democracies in Latin America.
Over the years, LAPOP studies have consistently demonstrated that the levels of
support for the system vary considerably from one country to the next but tend to remain
relatively stable within the countries. That is, countries with higher levels of system support
tend to maintain these levels over time, and low levels of support for the system tend to remain
low. A relatively rare exception to this pattern can take place when there are sudden or
extreme changes in the political context of a country, such as an economic crisis, a civil war,
or intense ethnic conflicts. Given that support for the system is measured individually, it is a
measure that is susceptible to the influence of important or powerful events in the political
circumstance. Therefore, under some circumstances, major shifts in the level of system support
for the system will occur.
The following figure illustrates the distribution of support for the system in 19 Latin
American and Caribbean countries measured by the AmericasBarometer in 2006-07. In the
figure it can be observed that countries such as Ecuador, Paraguay, Haiti, Brazil, Jamaica and
Peru have low levels of support for the system whereas Costa Rica, Mexico and Colombia
show higher levels of support.
75
80
60
System support
40
69.2
66.6
64.3 64.0
60.8
57.6 57.0 57.0
55.4 55.0
53.2 52.7 52.2 51.5
48.9
46.6 45.3
44.6 43.9
41.6
20 39.1 37.4
0 Ca U U Co M Do Ve C El H C G G Bo Ja Pa N Br Pe H Pa Ec
na nit ru
gu st ex m ne olo Sa ond hile uya uat li m na ica as ru ai
ti ra ua
da ed
a a ic
o
in zu m lv ur na em via ai
c m ra il gu do
St y R i c el bia a as a a g ay r
ic a a dor al ua
at a n a
es Re
pu
bl
ic
Source: Americas Barometer

2006 by LAPOP
Error bars: 95% CI
Figure IV-1. System Support: Per Country Averages. Comparative Sample, 2006-2007.
Beyond what the data says about each country specifically, on the whole they seem to
indicate that the most prominent changes in the way the State is managed are occurring in
countries where citizens have consistently declared themselves to be dissatisfied with the
performance of democratic governments. Thus, in this chapter we are concerned with
understanding the factors that reduce or increase the levels of support to the system and the
conditions under which these changes occur.
In 2004, Russel Dalton published a study on the erosion of system support in advanced
industrialized democracies in which he analyzed the factors that generated change in support
to the system. Several other studies (Easton 1976; Weatherford 1992; Anderson and Guillory
1997) have determined that the individual assessment of government performance is one of the
most important factors that influences the level of system support.
Dalton finds that, contrary to other studies, in highly industrialized countries the
individual assessment of the governments economic performance does not directly cause a
76
decrease in support for the system since citizens tend to have long-term vision regarding
economic policies. Following Daltons line of reasoning, the rest of this chapter will analyze
the degree to which individual evaluations of the economic performance of the governments in
Latin American countries determine the levels of support for the system. The principal
assumption in these examinations is that support for the system is strongly influenced by the
satisfaction or dissatisfaction with the domestic economic situation, resulting from government
performance in matters of economic policies.
The study will test two hypotheses which represent two distinct ways in which
economic factors can affect support for the system: the first is the perception of the national
and personal economic situations. The second test will analyze the way in which growth or
reduction in the level of development of a country can affect support for the system of a
specific country. In these tests, the level of development is measured by the Human
Development Index which the United Nations Development Program (UNDP) generates every
two years. The model will also control the influence of a countrys level of wealth, measured
by the Gross National Product (GNP), on the levels of support to the system. The basic
assumption of this second test holds that the most wealthy and most developed countries will
show greater levels of support for the system than poorer and less developed countries.
The combination of individual and aggregate factors enriches the understanding of the
phenomena studied, allowing for the combination of complementary elements which affect
political processes in different but simultaneous ways: the subject and the context.
The initial analysis of the determining factors for support for the system was carried
out through a linear regression which takes into consideration individual perceptions of the
national and personal economic situation and the perception of how generalized corruption is
among public officials, controlled by the sociodemographic characteristics of the citizens
interviewed. The analysis includes 19 countries from the AmericasBarometer 2006-07 in Latin
America and the Caribbean: Uruguay, Costa Rica, Mexico, Dominican Republic, Venezuela,
Colombia, El Salvador, Honduras, Chile, Guyana, Guatemala, Bolivia, Jamaica, Panama,
Nicaragua, Brazil, Peru, Haiti and Ecuador. Costa Rica is used as a reference category for the
analysis of the differences among the countries.
The results of this analysis, as seen in Table IV-1, demonstrate that contrary to what
happens in industrialized countries, in Latin America individual assessment of the national
economic situation and the personal economic situation do influence the decision to support
the political system or not. The results of the analysis of the regression also indicate that these
individual assessments and the perception of the level of generalized corruption are more
robust predictors of the level of support for the system for the countries analyzed, even more
important than the socioeconomic differences among individuals. In addition, the analysis
suggests that government performance in matters of economic policy is important to citizens in
Latin American and the Caribbean countries when determining their level of support for the
system.
Furthermore, the results of the linear regression suggest that there exists a difference in
support for the system which is determined by the country of residence of the people
77
interviewed. For example, Mexicans express greater levels of system support than Chileans,
Colombians or Bolivians. In other words, although there are differences among Mexicans
regarding their levels of support for the system, the support in Mexico is, on average, higher
than support in Chile, Colombia or Bolivia.
Table IV-1. Analysis of the Lineal Regression of Support for the System. Comparative Sample 2006-2007.
Source: Americas Barometer by LAPOP
Non-standardized Standardized
coefficients coefficients
Model B St. error Beta t Sig.
(Constant) 55.098 1.192 46.235 .000
Sex 1.187 .264 .026 4.499 .000
Age .844 .094 .055 8.941 .000
Education -.472 .201 -.016 -2.348 .019
Urban 1.985 .319 .040 6.225 .000
Wealth (individual
possession of material -.284 .083 -.025 -3.429 .001
goods)
National economic
.169 .007 .168 25.629 .000
situation
Individual economic
.097 .007 .087 13.266 .000
situation
Frequency of corruption
-.003 .000 -.106 -17.854 .000
acts
Mexico -3.468 .807 -.033 -4.296 .000
Guatemala -11.695 .846 -.107 -13.826 .000
Salvador -7.786 .814 -.077 -9.567 .000
Honduras -10.115 .817 -.099 -12.381 .000
Nicaragua -17.164 .838 -.167 -20.470 .000
Panama -18.018 .837 -.175 -21.518 .000
Colombia -7.972 .819 -.077 -9.739 .000
Ecuador -24.162 .806 -.240 -29.993 .000
Bolivia -14.737 .839 -.142 -17.565 .000
Peru -19.581 .832 -.194 -23.530 .000
Chile -14.235 .811 -.140 -17.561 .000
RDominicana -5.723 .815 -.056 -7.019 .000
Haiti -20.255 .875 -.194 -23.139 .000
Jamaica -13.796 .837 -.128 -16.483 .000
Guyana -13.871 .838 -.129 -16.552 .000
Uruguay -2.872 .815 -.028 -3.527 .000
Brazil -20.448 .806 -.200 -25.376 .000
Venezuela -7.697 .815 -.075 -9.441 .000
78
60
System support
40
61.5
55.2
48.8
44.2
20
0
Good Fair Bad Very bad
How would you qualify the national economic situation?
Source: Americas
Barometer 2006 by
LAPOP Error bars: 95% IC
Figure IV- 2. Support for the System per Perception of the National Economy. Comparative Sample
2006-2007.
It can be seen in Figure IV-2 that a favorable perception of the national economic
situation has a positive influence on the level of support that citizens give to the political
system. In general, people who perceive that their countrys economy is strong show higher
levels of support for the system than do persons who have a more unfavorable perception of
the national economic situation. The differences among those who think that the national
economic situation is good and those who think that it is neither good nor bad, that it is bad or
that it is very bad are statistically significant, as indicated by non-overlapping confidence
intervals.
The perception of ones personal economic situation shows the same relationship as the
tendency to support the system; people who are more satisfied with their personal economic
situation show greater levels of support for the system than those who consider that their
economic situation is bad or very bad. These results, both the national and personal evaluations
of economic situation, are valid and statistically significant for the 19 countries included in the
analysis.
79
60
System support
40
60.8 58.7
52.3
48.4
44.3
20
0
Very good Good Fair Bad Very bad
How would you qualify your personal economic situation?
Source: Americas
Barometer 2006 by LAPOP
Error bars: 95% IC
Figure IV-3. Support for the System according to Personal Economic Situation. Comparative Sample
2006-2007.
On the other hand, the perception of corruption has a negative relationship to support
for the system. As can be observed in the following figure, people who perceive generalized
corruption among public officials present lower levels of support for the system than those
who consider that corruption is less widespread. In the final section of this volume, additional
detail is provided on the impact of corruption on support.
80
60
50
System support
40
30
55.6 57.0
47.9
20
10
0
Very generalized Somewhat generalized Little generalized
According to your own experience or what you have heard,

corruption among public officials is...?
Source: Americas
Barometer 2006 by LAPOP Error bars: 95% IC
Figure IV-4. Support for the System according to Perception of Corruption. Comparative Sample
2006-2007.
Regarding the socioeconomic characteristics of respondents, the sex, age, education,

and level of personal wealth, all mark differences in the level of support that citizens show.
The results indicate that people who live in rural areas tend to support the system more than
those who live in urban areas, and older people support the system more than the young.
81
The following figure illustrates the way in which the level of personal wealth8, measured
according to capital goods owned, influences system support.
56
54
System support
52
50
0 1 2 3 4 5 6 7 8 9
Individual wealth measured by posession of capital goods

Source: Americas
Barmetro 2006 by
LAPOP
Figure IV-5. Support for the system according to level of personal wealth. Comparative Sample
2006-2007.
It can be seen in the figure that as the level of personal wealth increases, so does the
level of support for the system expressed by interviewees. Nevertheless, the increase in
support is not the same for all groups; rather it seems to be more moderate among people with
the lowest level of wealth (on the scale from points 0 to 2) and is more pronounced among
people with a medium-low to medium-high level of wealth (on the scale from points 2 to 7). It
is also interesting to observe that this relationship does not happen among those with most
personal wealth, where the level of support for the system remains practically constant even
when the level of personal wealth increases. However, the level of support for the system in
8
The personal level of wealth is measured through an additive index that results from sum of capital goods
possessed by the interviewee according to her own statement. These goods vary from having access to basic
services such as indoor plumbing and a sewer system to owning a car, a cell phone and other such goods.
82
this latter group of respondents is generally greater than the level of support among those with
less personal wealth.
54
53
System support
52
51
50
None Basic High Superior
Level of education
Source: Americas
Barometer 2006 by
LAPOP
Figure IV-6. Support for the System According to Educational Level. Comparative Sample 2006-2007.
As for the level of education of the respondents, Figure IV-6 illustrates a negative
relationship between the educational level and support for the system. In other words, as the
level of education increases, support for the system declines. People with no formal education
are the ones that express higher levels of support for the system, whereas those with higher
education are those that least support the system.
83
The following figure presents the differences in support for the system between men and
women. Figure IV-7 shows that in general, women express a slightly higher level of support
than do men. This difference is statistically significantthe confidence intervals do not
overlapalthough the differences are very small in substantive terms, since it is only one
point on a scale of 100.
60
50
System support
40
30
51.6 52.5
20
10
0
Male Female
Sex
Source: Americas
Barometer 2006 by
LAPOP
Error bars: 95% IC
Figure IV-7. Support for the System According to Gender. Comparative Sample 2006-2007.
Development, Economic Growth and Support for the System
In the previous section, it was seen that there are significant differences in support for
the system across countries. These differences cannot be attributed to any specific
phenomenon in each country in an analysis with data at the individual level, as we have done
until now, since there may be factors at the country level that cannot be detected at the
individual level, for example the influence of the political culture, the economic situation or
the political context, among many others.
84
However, the fact that the aspects that generate these differences cannot be detected
with data on the individual level does not mean that these differences are not real. In order to
determine which factor could generate these differences among countries, we will use an
analysis that combines information at the individual level with aggregate national level data.
Given that the development and economic growth of a country are, to a large extent,
the direct result, of the economic policies of the national government, the effect these have on
support for the system is, in effect, an evaluation of government performance in the economic
arena. In this section of the chapter, therefore, we concentrate on estimating the influence of
economic factors on support for the system. The following analysis will estimate in two
separate models the way in which the level of development in a country, measured in terms of
human development9, and the level of economic growth measured in terms of Gross Domestic
Product10, influence the different levels of support for the system among countries registered in
the previous section.
Following are the results of the analysis of mixed effects of economic growth on
support for the system.
Table IV-2. Analysis of Mixed Effects of Economic Growth on Support for the
System. Comparative Sample 2006-2007.
Variables Coefficients Error standard z P>z
National Economic -.0032463 .0003465 -9.37 0.000
Situation
Personal Economic -.001304 .0001476 -8.84 0.000
Situation
GDP2004 -.0000229 9.84e-06 -2.33 0.020
Constant 1.799696 .0638379 28.19 0.000
Mixed-effects REML regression # of obs = 31626
Group variable: country # of groups = 20
Prob > chi2 = 0.0000
The results presented in Table IV-2 indicate that, as expected, the differences in
national growthmeasured by GDPare statistically significant in the model and that the
individual perception of the national and personal economic situation continues to be
significant.
This means that the reasons, for example, for which Mexicans report higher levels of
system support than do Colombians, Bolivians, or Chileans is related to individual level
factors such as satisfaction with the national and personal economic situationsas mush as
with contextual economic factors at the national level. That is to say that the level of economic
growth in Mexico contributes to Mexicans showing higher levels of support because they think
the economy is healthy and because they feel that they live in a more developed country than
some of he others, which makes them report higher levels of system support than citizens of
9
The values for the Human Development Index for each country in the sample come from the 2004 estimation,
the most recent available. Source: Global Report on Human Development 2006, UNDP.
10
GDP values for each country are available only until 2004 and were taken from the Global Report on Human
Development 2006, UNDP.
85
countries.
Table IV-3. Analysis of Mixed Effects of level of Human Development

on Support for the System. Comparative Sample 2006-2007.
Variables Coefficients Error standard z P>z
Situacin econmica -.0032506 .0003471 -9.37 0.000
nacional
Situacin econmica -.0013032 .0001476 -8.83 0.000
personal
IDH2004 -.3828372 .3372549 -1.14 0.256
Cons 1.950214 .2541186 7.67 0.000
Mixed-effects REML regression # of obs = 31626
Group variable: country # of groups = 20
Prob > chi2 = 0.0000
On the other hand, differences among countries in the level of human development are
not statistically significant in the model, meaning that the national level of human development
in any given country does not influence the reported level of system support in different
countries.
These results provide persuasive evidence that individual political attitudes towards the
political system are not only influenced by individual characteristics but that they are also
influenced by structural elements that are part of the context in which the relations between
individual and political system take place.
Conclusions
In this chapter the evidence has shown that the study of support for the system can be
an important instrument for the evaluation of the stability of democracies. Evidence gathered
from the AmericasBarometer in 19 countries in Latin America and the Caribbean for this
analysis found that system support does not depend solely on political considerations; rather, it
is influenced in an important way by individual evaluations of government economic
performance, both in the perception of the national economic situation and in the perception of
citizens personal economic situation, contrary to what Dalton had found in highly
industrialized nations.
The study also found that the perception of generalized of corruption can diminish
support that citizens give to the system, reducing the levels of support proportionately to the
extent of such corruption.
These elements suggest that the countries where citizens perceive that governmental
economic performance generates economic problems for the country or poor conditions for the
individual economy of citizens will have lower levels of support for the system than those
countries in which the governments are able to maintain a healthy economy or at least a stable
one.
The evidence also suggests that there are differences between the levels of support for
86
the system among countries that are determined at least in part by the specific characteristics
of each country and that cannot be detected in the data at the individual level alone. The results
of the mixed-model analysis allows us to see that these differences are also the product of
differences in structural elements specific of the political system of each country, especially
the levels of economic growth which influence the individual levels of system support.
87
References
Anderson, C. J. and C. A. Guillory. "Political Institutions and Satisfaction with Democracy: A

Cross-National Analysis of Consensus and Majoritarian Systems." The American Political
Science Review 91 (1) (1997): 66-81.
Bastian, S. and R. Luckham. Introduction. Can Democracy Be Designed? Can Democracy Be
Designed? S. Bastian and R. Luckham. New York, Zed Books, 2003.
Dalton, R. J. Democratic Challenges, Democratic Choices: The Erosion of Political Support in
Advanced Industrial Democracies. Oxford, Oxford University Press, 2004.
Easton, D. . "A Re-Assessment of the Concept of Political Support." British Journal of Political
Science 5 (1975): 435-57.
Easton, D. "Theoretical Approaches to Political Support." Canadian Journal of Political Science
/ Revue canadienne de science politique 9 (3) (1976): 431 - 448.
Hetherington, M. J. "The Political Relevance of Political Trust." American Political Science
Review 92 (4) (1998): 791-808.
Huntington, S. The Third Wave: Democratization in the Late Twentieth Century. Norman,
University of Oklahoma Press, 1991.
Levi, M. and L. Stoker. "Political Trust and Trustworthiness." Annual Review of Political Science
3 (2000).
Miller, A H. "Political Issues and Trust in Government." American Political Science Review 68
(1974): 951-72.
Schwarz-Blum, V. Confianza en las instituciones: por que es necesaria? . D. Moreno.
Cochabamba, USAID Bolivia, 2006.
Seligson, M. A. "The Impact of Corruption on Regime Legitimacy: A Comparative Study of
Four Latin American Countries." Journal of Politics 64(2) (2002).
Weatherford, M. S. "Measuring Political Legitimacy." American Political Science Review 86(1)
(1992): 149-166.
88
V. Decentralize or Centralize? Challenges for Reform of the

State and Democracy in Latin America and the Caribbean1
Daniel Montalvo
Abstract
Since the beginning of the third wave of democratization, various debates have centered on
determining how to face the challenges that persist in the democracies of Latin America and the
Caribbean. A prevalent alternative in the region is decentralization, applied as a mechanism to
foster the vertical balance of powers of the state. However, upon exploring citizen public opinion
regarding political trust in local governments and the potential for lending support to a process
of decentralization, this chapter shows that there is considerable variation in trust in
municipalities and, contrary to expectations, up to 52 percent of people surveyed would like to
see the national government take over municipal obligations and services. In order to determine
the variables that explain these phenomena, in this research (1) a comparison is made of the
individuals political trust at different levels of the government; (2) political trust is studied as a
cause for support of a process of state reform; and (3) citizen support for institutional
decentralization or centralization is determined. This comparative analysis of citizen perception
is conducted in 19 countries, and individual case studies are carried out on extreme cases.
Numerous social scientists have debated the challenges that face democracy and
democratization worldwide. For instance, Dahl (1971) considers that it is practically beyond
discussion that the greater the socioeconomic level of a country, the greater the possibility for
that country to consolidate its democracy. Likewise, Sorensen (1998) attributes low levels of
education, urbanization and development to weak support for democracy. Other factors such as
corruption, adverse political culture, and social inequality are risks for a consolidated democracy
(Epstein et al., 2006). Over the last two decades, several countries have attempted to decentralize
their governments as a means to cope with the aforementioned challenges and, in this manner, to
increase the levels of democratic governance, to mitigate poverty, and to reduce corruption.
Tulia Faletti (2005) states that in fiscal terms, in 1980, local governments around the
world collected on average 15 percent of revenue and spent about 20 percent of [national]
expenditures, while at the end of the nineties these numbers expanded to 19 and 25 percent
respectively, and even doubled in some regions. Therefore Latin American and Caribbean
governments, with the support of several multinational organizations such as the World Bank, the
United Nations Development Program (UNDP), the German GTZ, and the United States Agency
1
This study has been possible thanks to USAID support of the AmericasBarometer. Additionally, our gratitude goes
to the USAID Mission in Ecuador, the Department of Political Science and its staff, the Center for the Americas, the
Provosts Office at the University of Vanderbilt. Special recognition goes to Professor Mitchell A. Seligson for his
valuable contributions to this study, and the team at the Latin American Public Opinion Project. Also thanks to
Professor Jonathan Hiskey for his extensive academic contributions.
89
for International Development (USAID), have extended their decentralization programs with
special emphasis on the region.2 In spite of the efforts that national governments and multilateral
organizations have made to solve different problems by way of decentralization, there is scarce
empirical evidence to determine whether these programs, from the point of view of the alleged
beneficiaries, have achieved their goals. To better understand the opinions of Latin American
and the Caribbean citizens regarding decentralization and centralization, the AmericasBarometer
in 2006-2007 included a block of questions that explore the attitudes of the regions inhabitants
regarding these processes of state reform.
This chapter analyzes fiscal and administrative centralization and decentralization from
the perspective of the inhabitants of 19 countries in Latin America and the Caribbean.3 It begins
with a conceptual description of centralization and decentralization, in conjunction with an
historical overview of the power transfer processes, examining the centrist tradition of the Latin
American and Caribbean states as an outcome of the colonial period. Next, theoretical arguments
are presented for the factors that might favor the processes of state reform, starting with political
trust in governmental institutions. Subsequently, countries are differentiated and described
according to those whose inhabitants have greater trust in their national governments and those
where municipalities are more trusted. Once political trust is determined, those who believe that
municipalities should have more obligations and receive more funding and those who, on the
contrary, support the national government assuming more obligations and municipal services, are
described. Lastly, countries whose population would be willing to pay higher taxes to the
municipality for better services are examined.
Decentralization, Centralization and Intergovernmental Transfers
Historical compilations of the processes of vertical balance of power in Latin America

and the Caribbean show the centralizing heritage received from the Spanish and Portuguese
colonizers in the Americas as an efficient means for extracting resources. During the 20th century
the central state in almost all Latin American countries successfully increased their levels of
power, authority and resource control. However, starting in the 1980s, concomitant with the
transitions to democracy, systematic processes of transfer of power from the national to the local
level were put in place (Seele 2004).
2
The UNDP area of Democratic Governance for example, contributes to the formation of national capacities; it also
fosters an adequate environment for effective decentralization, local governance and urban/rural development. The
technical area of Democracy and Governance at USAID directs its efforts at fighting corruption, promoting
democratic decentralization, legislative strengthening, civilian-military relations and effective implementation of
public policy. Lastly, the Decentralization and Subnational Thematic Group at the World Bank seeks to share and
deepen understanding regarding intergovernmental relationships, regional development and reduction of poverty, as
well as central and local governance to reinforce the effectiveness of the different levels of government.
3
The countries analyzed in this study, according to data availability, are: Mexico, Guatemala, Honduras, El
Salvador, Nicaragua, Costa Rica, Panama, Venezuela, Colombia, Ecuador, Brazil, Peru, Paraguay, Uruguay, Chile,
the Dominican Republic, Haiti, Jamaica and Guyana. The data was obtained from the 2006 Round of surveys
conducted by the AmericasBarometer and the Latin American Public Opinion Project (LAPOP).
90
What is the theory behind decentralization and centralization in the modern nation-states
of Latin America and the Caribbean? In political science terms, Tulia Falleti (2005) calls
decentralization the process of state reform through public policies, focused on the transfer of
power, responsibilities and/or resources from higher levels to lower levels of government. On the
other hand, centralization is the process of state reform through public policies that are aimed at
the restitution at higher levels of government of those faculties attributed to lower levels in a
determined political context. When analyzing the components of these definitions it can be seen
that primary emphasis is given to the dynamic character of these processes, rather than to the
static characteristics of a system. For this reason, the character of state reform is studied and not
an institutional description of a specific moment in time. The character of state reform excludes
the possibility of privatization of public functions, which means that the transfer of faculties is
strictly within the public sector. According to the political constitution of the states analyzed in
this study, several levels of government are contemplated: (1) the national level of government,
which corresponds to the national government or central government4, (2) the intermediate level
of government, that corresponds to the government of federated states, departments or provinces,
(3) the local level of government, that corresponds to municipalities or city councils, and (4)
other subnational levels or jurisdictions.5
The transfers that take place between different levels of government occur in three areas:
(1) The Political area refers to granting space for territorial representation on the basis of
electoral reforms regarding the determination of executive and legislative branches of
government. Additionally, it refers to the transfer of authority to designate and remove personnel
in the public sector at different governmental levels. (2) The Fiscal area refers to the ability to
collect and distribute revenue in the form of taxes, carry out income and transfers between the
different levels of government, distribute expenditures as expenses and investment, and the
authority to manage public debt; and (3) The Administrative area, which refers to the transference
of authority to grant public services (Manor 1999; Tulchin 2004; ONeill 2005; Falleti 2005).
Falleti (2005) identifies the sequence in which political, fiscal and administrative transference
takes place. When authority is transferred, the preferred order is first, the transfer of
administrative control and the relinquishing of responsibilities for public services; then fiscal
transfer retaining political control and power. Conversely, the level of government to which
authority is transferred should receive an initial transference of political control and power in
order to avoid being designated or removed at the discretion of the other level of government.
Once its legitimacy is insured by vote, it requires fiscal control to consolidate its institutional
operation and finally exert administrative control of public services.
To better explain the processes of decentralization that have taken place during the third
wave of democratization in Latin America, a number of social scientists have restated the
definition in more precise terms. Kathleen ONeil (2005) defines decentralization in terms of (1)
devolution, the process of transfer of funds and responsibilities to officials at the intermediate and
local levels, designated by the national government; (2) delegation, the process of conferring
responsibilities to non-official organizations6 and (3) de-concentration, the process of conferring
4
For standardization purposes, in this study the term national government is used in reference to the central
government.
5
For example, counties, villages, parochial boards, and so on.
6
The delegation is done by means of privatization or concession of public works.
91
decision-making power to a series of ministers in the executive branch. A distinction also must be
made between federalism, which corresponds to a system of government in which sovereignty is
constitutionally divided between the authority of the national government and subnational
political units, and in which power is shared between the national government and intermediate
governments7, and autonomy, in which subnational levels of government have ample executive
and legislative powers. Sovereignty as an attribute, however, belongs exclusively to the nation.
The State of Decentralization in Latin America and the Caribbean
Several researchers agree that the consolidation of state reforms such as political, fiscal
and administrative transfers entered into effect with the return of democracy to Latin American
and the Caribbean at the end of the 1970s.8 Daughters and Harper (2007) argue that early state-
decentralization reforms took place in the political arena, specifically with the creation of local-
level representative democracies. Mayoral elections began with high expectations, given the
natural advantage the local and intermediate levels have over the national level to link the needs
of the population with the goods and services offered by local governments. Furthermore, another
advantage lies in the facility of interpreting preferences through an increase in political
participation in the community. In the mid 1990s, one of the first conditions for political
decentralization was widely accomplished among the nations of Latin America and the
Caribbeandemocratic mayoral elections were instituted in the entire region, except in
Surinam.9 The first country to put elections for local government into effect was Mexico in 1917.
This coincided with the year of its return to democracy. The last countries in the region to carry
out local elections were Guyana and Panama, both in 1995, and the years of their transition to
democracy were 1966 and 1989 respectively. In Table V-I the first years of mayor elections in
the countries are compared, along with the years of democratic transition.
Regarding state fiscal policy reform, subnational governments have made use of three
general forms of financing: (1) excise taxes,10 (2) intergovernmental transfers,11 and (3) the
generation of income by looking for sources of financing themselves.12 Falleti (2005) says that in
the 1980s, subnational governments collected, on average, 15% of total income, and spent on
average 20% of total expenditures. This data according to Daughters and Harper (2007) hides a
significant difference among countries, particularly in terms of expenses and investment. In the
period from 1996 to 2004, Inter-American Development Bank data reveals that, in descending
order, three countriesArgentina, Brazil and Colombiahad high rates of decentralization of
expenses with percentages close to 50%, placing them among the group of the most decentralized
countries in the world.13 A second group of countriesMexico Venezuela, Bolivia, Peru and
7
Generally states or provinces.
8
See Robert Daughters and Leslie Harper (2007), Tulia Falleti (2005), Kathleen ONeill (2005) Daniel Treisman
(2002).
9
Cuba is not taken into account in this analysis.
10
The most common are taxes on property, vehicles, and industry and commerce.
11
They can be made permanent and egalitarian through legislation or at the discretion of higher levels of
government.
12
This is in the case of municipal water and power companies, etc.
13
The list of these countries includes Canada, the United State, and North-European countries (Daughters y Harper,
2007).
92
Ecuadorpresent a moderate yet equally significant level of expenditure decentralization,

between 17.5% and 31.8% in 2004. It is worth mentioning that, out of the first five countries,
four are federal systems of government14, except for Colombia which has had an historically
regionalist tradition that has allowed for higher levels of decentralization.
A third group of countriesUruguay, Chile, Honduras, Guatemala, El Salvador,

Paraguay, Nicaragua, Costa Rica and Panamaare characterized by low levels of expenditure
decentralization, from 13% in the cases of Uruguay and Chile, to 1% and 3% in the cases of
Panama and Costa Rica.
In the same vein, since the mid 1990s, a growing number of restrictions on subnational
indebtedness have been put in place in order to avoid the problems of excessive debt that arose
from the policies that had been adopted by Brazil and Colombia in the first instance, followed by
Bolivia, Ecuador and Peru. A summary of the characteristics of subnational public indebtedness
is presented in Figure V.1.
14
Argentina, Brazil, Mexico and Venezuela (Central Intelligence Agency, 2007).
93
Table V.1. First year municipal elections, democratic transition and subnational indebtedness
Restriction
Authorization
Municipal Year of Prohibition of in the Use
from National
Country Mayor Transition to Subnational of Public
Government for
Elections Democracy* Indebtedness** Debt
Indebtedness**
Funding**
Argentina 1983 1983 9 9
Belize 1981 1981 n.a. n.a. n.a.
Bolivia 1985 1982 9 9
Brazil n.a. a 1985 9 9
Chile 1992 1990 9
Colombia 1988 1958 9 9
Costa Rica 1949 1949 9 9
Dominican
Republic 1966 1966 n.a. n.a. n.a.
Ecuador 1983 1979 9 9
El Salvador 1985 1984 9 9
Guatemala 1985 1985 9 9
Guyana 1995 1966 9 9
Honduras 1982 1982 9 9
Jamaica 1962 1962 9
Mexico 1917 1917 9 9
Nicaragua 1992 1990 n.a. n.a. n.a.
Panama 1995 1989 9 9
Paraguay 1991 1989 9 9
Peru 1980 1980 9 9
b
Surinam n.e. 1987 9
Trinidad y
Tobago 1962 1962 9 9
Uruguay 1984 1985 n.a. n.a. n.a.
Venezuela 1989 1958 n.a. n.a. n.a.
Sources: * Inter American Development Bank (1997).
** Ibid (2004) Created by Robert Daughters and Leslie Harper (2007) and restructured by the author.
a
n.a. Data not available b Local government is non-existent.
Public Opinion: Political Trust among Institutions and Support for

Decentralization
What do the people of Latin America and the Caribbean think about the processes of
decentralization? Do they trust national governments more, or does their trust rest more in the
local levels of government? Would they be willing to back municipalities receiving more
94
responsibilities and resources from the national government, or do they think that the national
government needs to assume more municipal responsibilities and resources? There have been
many attempts at defining the ideal institutional balance between centralization and
decentralization of the state, but very few studies have looked into public opinion regarding these
state-reform processes. First of all, this study analyzes the level of political trust that people in
the region have in their public institutions, focusing principally on the national government and
the municipality. According to Hetherington (2005) political trust is the degree of perception that
people have of whether or not the government is producing results that are consistent with their
expectations. The key characteristic of this concept is the peoples perception of government
performance and not government performance in and of itself. Obviously perceptions must
connect with reality, but often the link is ambiguous and indirect. This leads to several problems,
not the least of which is the difficulty for citizens to determine what their governments
performance really is. Much of this information is transmitted by the media, and perceptions can
be distorted for political purposes. Moreover, access to the media information varies according to
socioeconomic status and the citizens level of interest in politics.
Trust in the institutions is related to support for democracy and other democratic
principles. Adrian and Smith (2006) point out that in both democratic systems and in nations in
the process of democratization, individuals who assess present and past regimes unfavorably also
express very little trust in institutions. Support is defined as a foundation, confirmation or proof
of an opinion or doctrine. Given the premise that political trust is, ultimately, related to support
for democracy and democratic processes, it can be inferred that the greater the level of trust, for
example, in the municipality, the greater the support for decentralization. This study does not
analyze the inverse relationship, which is that support for decentralization causes individuals to
have greater trust in the institutions because decentralization is seen from a dynamic perspective.
People decide whether they will favor a greater decentralization based on personal experience
with municipal government vis--vis their experience with the national government.
In this chapter I analyze data obtained in the 2006-07 round of surveys carried out by the
Latin American Public Opinion Project (LAPOP) as part of the AmericasBarometer series.
Specifically the main items used are:
LGL2. In your opinion, should more obligations and more funding be given to the municipality, or
should the national government be allowed to assume more municipal obligations and services?
B32. To what extent do you trust your municipality?
95
Figure V.1. Trust in the municipality per support for decentralization
The strong result of the connection between citizen support for decentralization and trust
in the municipality is shown in Figure V.1. In the nineteen countries analyzed,15 the greater the
political trust in the municipality, the greater the support for decentralization. This finding
verifies the hypothesis that, as for individuals, an increase in trust in the municipality means there
is greater support for a process of decentralization. The latter indicates that in terms of public
policy, it is of fundamental importance to take into account that citizens will support a process of
transference to municipalities when they perceive that their mayors, city council members and
municipal employees are performing well. That is to say, municipal officials have a fundamental
role to play in building the necessary citizen support in the establishment of a process of state-
decentralization because to the degree that municipalities fulfill individual expectations, the result
is directly translated into citizens favoring the increase of political and economic power for local
governments.16
15
According to the available information, the following countries were included in this study: Chile, Colombia,
Costa Rica, Ecuador, El Salvador, Guatemala, Guyana, Haiti, Honduras, Jamaica, Mexico, Nicaragua, Panama,
Paraguay, Peru and the Dominican Republic.
16
The relation between political Trust in the national government and support for centralization is not significant.
96
Comparative Level of Municipal Political Trust versus Other Democratic

Institutions
How much political trust do the people of Latin America and the Caribbean have in
municipalities? How does trust in municipalities differ from political trust in other state
institutions? By assessing the ten principal institutions that are present in a democratic regime,
the levels of political trust in the countries of Latin America and the Caribbean can be analyzed.
Figure V.2 summarizes the results of these questions in all nineteen countries. On a scale of 100
points, where 100 means the person has a high level of trust in the institution and zero means the
person has absolutely no Trust, Latin Americans and people in the Caribbean who responded to
this survey gave the municipality a score of about fifty points, whereas the national government
got 49 points. The average trust for all institutions is at 48 points, that is, trust in the municipality
is two points above the average for the rest of the institutions, whereas trust in the national
government is one point above the institutional average. It can be seen that the armed forces is
the institution in which the population places the highest level of Trust, as it is 13 points over the
institutional average. At the same time, political parties are the least trusted institutions, at 13
points below the average. On the graph, in black, there are the municipality and the national
government. It should be noted that the difference between the level of trust in municipalities and
national governments in the 19 countries is minimal (less than one point).
Figure V.2. Political trust in public institutions.
97
Political Trust in Municipal Government per Country
Are these levels of trust in municipalities the same in all countries? In this section a
comparative analysis is made of the level of political trust citizens have in their municipalities by
country of residence. In Figure V.3 it can be seen that on a scale of zero to 100 points, where 100
means a lot of trust and zero means none at all, at 60 points, Dominican citizens are the ones who
have the highest level of political trust in the municipalities, awarding them a score of 60 points.
This is to say that, in the Dominican Republic, trust in the municipality in is ten points above the
average political trust level for the municipalities of Latin America and the Caribbean, which is
50 points. Haiti is in last place with 30 points for trust in municipalities, 20 points below the
average for the municipalities in the region. Furthermore, there exists a group of countries
Nicaragua, Guyana, Colombia and Brazil that show a level of trust that is very close to the
regional average of political trust in municipalities.
70
60
Trust in the Municipality
Mean
50
40
.
30 60.0 59.6 58.5

56.5 56.4 55.5
53.7
51.6
50.2 49.8 49.1 48.6 47.8 47.3
45.3
43.2
41.7 41.6
20
29.6
10
0
D
El inic
U lva Re
M ua y r
G co
H em
C ur
N
G rag
C na
B mb
C il
Pa a R
Ec gua
Pa do
Ve m a
Ja zue
Pe ica
H
om
ru d
ol
ra ia
os
ai
on a
hi as
ic
ua
uy ua
ex
m la
Sa an
ra ica
ua y
na r
ne
ru
ti
a
le
z
g o p.
t
d la
i
t
Source: AmericasBarometer 2006

Error Bars: 95% CI
Figure V.3. Political Trust in Municipalities per country
98
Political Trust in the Municipal Government Compared with Trust in the National
Government, per Country
Despite many shared characteristics among the various Latin American and Caribbean
countries studied, there are key differences in trust in municipalities and national governments.
The data reveal three groupings: (1) Countries where political trust in the municipality is greater
than that in the national government; (2) countries in which political trust for the national
government is greater than the trust in the municipality; and (3) countries where the difference in
the political trust in the national government and the municipality is not significant. Figure V.4
illustrates the three groups where there is a difference between political trust in municipalities
and national government. The bars with positive scores show countries where there is more trust
in the municipality, whereas bars with negative scores show greater trust in the national
government.
30
Trust Difference Between the Municipality
25.6
25
and the National Government
20
14.3
13.2
15 12.5
10 7.0
4.0 3.3
5
-5 -1.1 -1.1 -1.6 -1.7 -2.5

-4.7
-10 -5.3 -5.8 -6.0
-8.5
-15 -11.3
-20
-25
-23.8
-30
Ec
Ni do
Pa ag
G gua
El em
Pe lva
H
Ja ur
B ica
Pa il
G ma
M na
Co co
D aR
C inic
U
Co ay
Ve b
Ha zue
on
ra
om ic
hi
ru
ua y
uy
ex
ca r
m as
ua
ra ua
Sa ala
ru do
na
ne ia
iti la
st
lo
le n
s
gu
t
a
i
a
r
m
a
a
r
Re
pu
bl
ic
Error Bars: 95% CI
Figure V.4. Political trust in the national government versus municipality
99
It is interesting to note that the country with the highest dispersion range between political
trust in the municipality and the national government is Ecuador with almost 26 points. Thus,
Ecuadorians show a higher degree of political trust in the municipality in comparison to the other
Latin American and Caribbean studied. An analysis of this case indicates that the results are
systematically related to the high level of political instability in Ecuadors national government in
the last twelve years. According to Seligson et al. (2006), between 1995 and 2007, Ecuador has
had eight presidents, each one holding office for an average of 1.5 years. The crisis at the national
level government of Ecuador began in 1995 with the exit from the country of then vice-president
Alberto Dahik, during the administration of Sixto Durn-Ballen, due to a political trial against
Dahik for misuse of reserve funds. In 1997, president-elect Abadala Bucarm, after becoming
involved in several corruption scandals, was removed from office through a National Congress
resolution based on a constitutional provision allowing for mental unfitness as grounds for
dismissal. Following an interim period presided over by former president of the National
Congress Fabian Alarcon, presidential elections were held in 1998, and former Quito Mayor
Jamil Mahuad took office. Due to a financial crisis that resulted in bank accounts being frozen,
Mahuad was forced to resign in 2001 for putting the country in the worst economic crisis since
the establishment of the republic. The vice-president, Gustavo Noboa succeeded him until former
military man Lucio Gutierrez, who had participated in the coup detat against Mahuad, came to
power after being elected president of the republic in 2002 (Seligson et al. 2004). In 2005,
Gutierrez and the National Congress ordered the removal from office of all the minister judges of
the Supreme Court of Justice, adducing high politicization and partiality in the Ecuadorian justice
system. These events in turn provoked an abrupt reaction from a movement in the capital, and the
National Congress was forced to remove Gutierrez from office by means of a constitutional
definition of abandonment of office. He was succeeded by his vice-president Alfredo Palacio,
and in 2006 yet another presidential election took place (Seligson et al., 2006). At the time of this
writing, current Ecuadorian president Rafael Correa recently held a national referendum to
authorize the rewriting of Ecuadors constitution for the twentieth time.
Political instability has not affected all levels of the government equally. For instance, the
prefect of the province of Guayas, Nicolas Lapenti, has been democratically re-elected for the
fourth time, and will have held office for 16 years. The municipal mayors of Quito and Guayaquil
both have mayors who have been re-elected and will have been in office eight years. For all these
reasons Ecuador presents itself as an extreme case in which political trust is much greater at the
local government level than that of the national government. The rest of the countries studied,
which have higher levels of political trust in municipalities than in the national government, show
a much more moderate difference. 17
In the second group of countries citizens express a higher level of institutional trust in the
national government compared with municipalities. Figure V.4 shows the eight countries that
share this characteristic in this study. In this figure, the citizens of countries below the horizontal
line trust the national government more than the local government. Haiti appears at the far end
with a much higher level of trust in the national government than in municipalities, with a
twenty-four point difference. An analysis of this case in the report Political Culture in Haiti:
2006 (Zephyr and Pierre, 2007) indicates that since Haitian independence in 1804, the country
17
Nicaragua 14,3; Paraguay 13,2; Guatemala12,5; El Salvador 7,0; Peru 4,0; y Honduras 3,3.
100
has been marked by a tradition of administrative and fiscal centralization. When Haiti held its
first elections for its Municipal Communal Councils in 1844, then president Riviere Herard
established a law creating a series of districts and prefectures under the control of the president in
order to counteract the municipal councils political power. Subsequent governments in Haiti
have employed municipal governments as extensions of central power, a process that was due to
constant electoral fraud. In response to this domination by the central government, the inhabitants
themselves created their own local governmental institutions which have been informally
legitimized.
Local community development associations were put in place during the Duvalier regime;
however, their functioning has been far from democratic (Zephyr and Pierre 2007). Called
Community Action Councils (CAC) these associations were increasingly pressured defer to
central power. Since the CACs were tied to Duvaliers power, these institutions ceased to exist
in 1986. With the advent of the new constitution in 1987, the formation of local institutions has
not been permitted. For example, Departmental Assemblies in which the members of the Inter-
departmental Council should be chosen were never held. It was thus impossible to establish
municipalities. Starting in 1990, Haitians supported candidates from different political parties for
the establishment of local governments. In 1995, local inhabitants motivated members of local
groups to run for local offices. This gave rise to an important stage in the development of local
governance; they did not limit themselves to supporting external political forces, rather they
undertook campaigns to participate in local elections. Unfortunately, Haitian Municipalities do
not have the logistical, technical or financial means to respond to the needs of their jurisdictions.
In most cases, mayors depend on foreign contributions to finance small development projects.
Generally such projects are financed and managed by NGOs or other international organizations
in collaboration with local associations and/or neighborhood committees, thus leaving out of the
equation the locally elected officials. Local level governmental projects are initiated sporadically,
although the law of April 4th, 1996 calls for persons elected to occupy local offices to fully
participate in local development projects. In the absence of effective municipal institutions,
Haitians trust the national government rather than local governments.
Regarding the rest of the countries in this category, in spite of the fact that citizens show
more trust in the national government than in municipalities, the gap is not as great as in the case
of Haiti.18 In Figure V.4 a third group of countries shows that the difference between political
trust in the national and municipal government is not significant. This is to say, that these
countries exhibit no statistical difference in their trust in government at the local or national
levels.19 In sum, it has been determined that on a scale from zero to 100 the inhabitants of the
countries in this study give an average trust score of 45 points to ten of the most representative
institutions of a democratic regime. When the countries are organized in descending order
according to the sum of political trust in the national government plus trust in the municipality,
the Dominican Republic is the country with the greatest level of trust in the two institutions,
18
Venezuela 11.3; Colombia 8,5; Uruguay 6.0; Chile 5,8; the Dominican Republic 5,3; Costa Rica 4,7 and Mexico
2,5.
19
Guyana, Panama, Brazil and Jamaica with a trust level of 95%. In Figure IV.4, it can be seen that the error bars
coincide with the horizontal line that indicates that there is no statistical difference in the level of Trust in both levels
of government.
101
followed by Mexico, Chile, El Salvador, Colombia, Honduras, Guyana, Costa Rica, Guatemala,
Panama, Nicaragua, Haiti, Jamaica, Paraguay, Peru and finally Ecuador.
Variables that Determine Political Trust in Municipalities in Latin America and the
Caribbean
This section reports on the factors that cause inhabitants to have greater trust at the local
level than at the national level of government. To this end, after having determined the group of
countries whose inhabitants have more trust in the municipalityParaguay, Peru, Ecuador,
Nicaragua, Honduras, El Salvador and Guatemalaa linear regression is done to determine some
of the variables that explain trust in municipalities. Kenneth Newton and Pippa Norris (1999)
identified at least three schools of thought regarding political trust in the institutions of the state,
with the following explanatory factors: (1) Those that focus on the individuals socio-
psychological characteristics; (2) those that focus on the cultural environment of individuals,
groups and communities; and (3) those that concentrate on the performance of government
institutions. 20
According to the socio-psychological school, Newton and Norris (1999) determined that
the inner quality of trust between people, alone or joined with optimism, make-up the basic
trust group that is formed in the first stages of psychological development as a result of the
mother-child feeding experience. In that study, the authors cite Easton (1965) who concluded that
the presence of trust means that the needs and interests of the members in a society are being
taken care of by the authorities even if there is little supervision or scrutiny. If this school of
thought is correct, it can be expected that there is an association between trust at the individual
level and trust in political institutions. In order to establish which factors determine the levels of
political trust in municipalities, the questions appearing in Appendix 1 were utilized. Another
school of thought considers that participation in community activities and socialization increase
general trust between a group of people and institutions, which means that the greater the social
capital, the greater the persons general level of Trust. Finally, Newton and Norris (1999)
consider that governmental institutions with good performance records generate greater trust in
people; conversely, inefficient institutions generate rejection.
A model was created in order to verify these theories, which, aside from including
explanatory variables from the three schools of thought mentioned above, was controlled by the
standard socioeconomic variables such as: (1) urban/rural dwelling, (2) gender, (3) age, (4)
educational level, (5) wealth measured by possession of capital goods and (6) dummy variables
to control the country-effects. The results of this regression can be seen in Appendix 2.
The results show that the levels of political trust in the municipality do not depend on the
control variables: gender or wealth measured by the possession of capital goods. What actually
20
As the format of the Paraguay questionnaire is different from the one used in the rest of the countries, due to the
fact that this study was not carried out by the LAPOP consortium, the results of some of the explanatory variables
form outliers in the sample, therefore, Paraguay was removed from the rest of this analysis.
102
determines the level of political trust in the municipality is the level of education. As the level of
education increases, political trust in municipalities decreases, probably because persons with
higher levels of education tend to be more aware of institutional failures. This is troubling in that
it is precisely the highly educated segment of the population that could accelerate the process of
state reform, although they are also the most critical of the system. In addition, these results
suggest that a populist candidate could obtain a great deal of benefit by pressuring the system to
implement a process of decentralization. In Table V.2, as follows, there is a summary of some of
the explanatory statistically significant variables and their relationpositive or negativeto
political trust in municipalities.
Table V.2. Predictors of political trust in municipalities

Variables Effect on political trust in the municipality
Age +
Education -
Urban sector +
Interpersonal trust +
Satisfaction with life +
Victim of corruption -
Perception of municipal services +
Attendance at town meetings +
Victim of a criminal act -
Table V.2 shows that as age increases and the individual resides in more urbanized areas,
trust in the municipality increases. Additionally, it is shown that the greater the degree of
interpersonal Trust, the greater the intensity of political trust in municipalities. This positive
relation is also expressed in satisfaction with life. In the area of institutional performance
according to individual requirements, people who have been victims of corruption at the level of
public institutions in the last year, trust the municipality less.21 Once again this confirms the
importance of the effect of the perception of corruption on individual trust in municipalities.
While it is not surprising that people distrust institutions when they have been victims of
corruption in their interactions with them, an important finding of this study is that individuals
can also generalize their distrust to include other state institutions. In other words, if an individual
is a victim of corruption in national institutions, the same individual might decide that all state
institutions are untrustworthy, including those at the local level. The same is true with victims of
crime, as these people also show distrust in the municipality.
Citizens who perceive the municipality as a purveyor of services and fair treatment, show
a greater level of political trust in municipalities. This is a fundamental finding in terms of a
decentralization process. It is clear that in order to generate political trust and the subsequent
21
In the regression, a variable was included to measure the degree of municipal-level corruption but it turned out to
be non significant.
103
support for decentralization, mayors, city council members and other municipal officials must
optimize municipal services and pay sustained and timely attention to citizen demands. This
process of trust building is vital prior to a decentralization process as will be shown in later
sections.
Centralize or decentralize?
Having described the reasons why people feel political trust or distrust in the
municipality, it is now to be determined whether Latin American and Caribbean people in the
nineteen countries included in this study want municipalities to undertake more responsibilities
and receive more money from the national government, or if, on the contrary, they think the
national government should assume the municipalities responsibilities and administer their
services. In order to determine the level of support for fiscal and administrative decentralization
or centralization the following question was asked: 22
LGL2. In your opinion, should more obligations and more money be assigned to the municipality, or
should the national government be allowed to assume more municipal obligations and services?
The results from this question are depicted in Figure V.5.
22
In this study there is no analysis of public opinion regarding political decentralization and centralization
104
National
Government
should acquire
more
responsibilities
and municipal
services
More money
and
48.3% 51.7% responsibilities
should be given
to the
municipality
Figure V.5. Percentage of support for state decentralization or centralization
The graph shows that the opinions of the inhabitants of the Latin American and Caribbean
countries that participated in this study are closely divided between centralization and
decentralization of the state. This result requires careful analysis of the countries in which people
think there should be a greater process of decentralization, as well as those countries where
people think that the national government should assume the municipal functions. In figure V.6
Jamaica appears first among the countries whose citizens prefer a greater fiscal and
administrative transference; more than twice as many people in the population prefer a process of
decentralization to one of centralization.
105
100%
Centralizar
Descentralizar
32.1%
80% 38.4% 40.2% 42.5% 42.9% 45.1% 45.2%
Support
60%
40%
67.9%
61.6% 59.8% 57.5% 57.1% 54.9% 54.8%
20%
0%
Ja
Ec
ic
om
hi
uy
ua
m
ua
ar
le
an
te
ai
in
do
ag
ca
m
a
ic
r
u
al
an
a
R
ep
.
Figure V.6. Countries with more support for decentralization
This may be due in part to the fact that in early 2005 several local Jamaican authorities had
prepared Local Sustainable Development Plans. In the case of Kingston and Saint Andrew
Corporation, the most populous areas of Jamaica, these plans have benefited from World Bank
collaboration through financial management training and the training of Parochial Board
Members in public administration. 23 The local government system in Jamaica was set up by the
British in the second half of the eighteenth century. In 1944, the system of parochial boards
was established in order to increase political representation at the local level. This system of local
representation is still in place today in Jamaica. Currently, Jamaicans perceive the most serious
problems the country faces to be the following: deterioration in the roadway system, water
shortage and poor sewer system service. These problems correspond to the need to assign more
responsibilities and resources to local governments so they can repair infrastructure and improve
basic services.
In the rest of the countriesNicaragua, Guyana, Guatemala, the Dominican Republic,

Chile and Ecuadorover 50% of persons surveyed were in favor of transferring more
23
Parochial Board Members are elected by popular vote at the local level.
106
responsibilities and resources to the municipalities. However, there is a second group of countries
in which interviewees considered that the national government should be the one to assume more
obligations and municipal services. The results regarding administrative and fiscal centralization
are summarized in Figure V.7.
100% Decentralize
Centralize
21.8%
34.9%
37.2% 38.2% 39.0% 39.1%
80%
43.8% 45.4% 46.2% 47.2% 48.7%
60%
Support
40% 78.2%
65.1%
62.8% 61.8% 61.0% 60.9%
56.2% 54.6% 53.8% 52.8% 51.3%
20%
0%
P
C
an
lS
er
ol
ra
en
os
on
ai
ru
ex
om
ti
u
zi
gu
am
ta
ez
al
du
ic
l
va
o
ay
ue
bi
ra
R
a
ic
do
a
la
a
r
Figure V.7. Countries with greater support for centralization
It can be seen that Panamanians who consider that the national government ought to
assume more municipal obligations and services are 3.5 times more numerous than those who
favor decentralization.. These results are similar to those of Daughters and Harper (2007) who
found Panama to be the least decentralized country in the region.24 In the study for the BID
Municipal Development and Decentralization Support Program, Nessim et al. (2003) stated that
tax revenue for Panamanian municipalities constitutes taxes on economic activity, vehicle
circulation tax, and construction permits. Property tax is not a municipal tax. This means that at
the national level, two out of every three municipalities do not have enough resources to cover
operating costs. For this reason, the quality of municipal services is poor and in general produces
an unfavorable opinion of municipal governments. The difference is less in the rest of the
24
In fiscal and administrative terms; in terms of political decentralization, along with Guyana, Panama has been one
of the last countries to implement local elections for municipal mayors.
107
countriesColombia, Honduras, Haiti, El Salvador, Peru, Mxico and Costa Ricain spite of
the fact that they belong to the group of countries with greater support for centralization than
decentralization of the state.
The battery of questions used to explain some of the explanatory variables for support for
centralization or decentralization of the state appears in Appendix 3. Also, the following socio-
demographic control variables were included: (1) urban/rural dwelling; (2) gender; (3) age; (4)
educational level; (5) wealth measured by possession of capital goods; and (6) dummy variables
to control the country-effects. The results of the binomial logistic regression can be seen in
Appendix 4. Table V.3, sums up some of the statistically significant explanatory variables and
their relation to the centralization or decentralization of the state.
Table V.3. Predictors for Decentralization and Centralization
Effect on decentralization or centralization

Variables
of the state
Sex (Male) Decentralization
Age Decentralization
Urban sector Decentralization
Level of education Decentralization
Incumbent governments efficiency Centralization
Political knowledge Decentralization
Perception of Municipal Services Decentralization
Trust in the municipality Decentralization
Interest in politics Centralization
Table V.3 shows that in socio-demographic terms, sex, age, educational level and
the geographic area where Latin Americans and Caribbean people live, determine their support
for decentralization. Therefore, men, persons who are older or possess a higher educational level
and residents of urban areas, are more in favor of fiscal and administrative decentralization.
Likewise, people with more political knowledge, who have requested municipal personnel
assistance or who perceive that municipalities treat citizens fairly and fulfill their needs, also tend
to exhibit greater support for descentralization. Nevertheless, people who have in interest in
108
politics and consider that the incumbent national government is efficient, show greater support
for centralization. This could be because people who are interested in politics seek public office
in the national level of the government, and for this reason they do not choose decentralization as
it could take power from the national government.
Finally, interviewees were asked, Would you be willing to pay higher taxes to the
municipality in order for it to be able to give better municipal services or do you believe it is not
worthwhile to pay more taxes to the municipality? Except in Haiti, where eighty percent of
those interviewed said they would be willing to pay higher taxes, more than sixty percent of
people in the nineteen countries analyzed in this study answered that it was not worthwhile to pay
more taxes, The refusal to pay more taxes is understandable, since people do not, as a rule,
volunteer to pay taxes.
Conclusion
This chapter has analyzed public opinion regarding financial and administrative
centralization and decentralization state reforms in Latin America and the Caribbean. In spite of
enormous efforts made for decentralization in the region, there are few empirical studies of the
populations opinion regarding these processes of reform. This analysis is an attempt to address
this deficit by means of a nineteen country survey in the region that makes possible the analysis
of levels of trust in the central government and municipalities and the support that the people of
Latin America and the Caribbean give the processes of decentralization and centralization.
Contrary to what might be expected regarding massive support for decentralization, the
population of Latin America and the Caribbean is divided in its opinion regarding this process.
Nearly half of the people think that more obligations and resources should be given to
municipalities, while the other half believe that it is the central government that should assume
municipal obligations and administer resources. This coincides with the minimal difference
between trust in the national and local governments. Therefore, it is necessary to determine the
support governments and international organizations and other cooperation organizations receive
from the population when planning to aid reform processes in the transference of resources and
responsibilities at an intergovernmental level.
A case by case analysis of the relation between trust in the municipality and the national
government, and the ensuing support for centralization or decentralization, has found that in
Ecuador, Guatemala and Nicaragua inhabitants trust their municipalities more, and would be
willing to support a decentralization process. In the case of Guyana and Jamaica, although the
level of trust is practically the same for national government and municipality, people in these
countries would be willing to support the process of decentralization. Finally, Chile and the
Dominican Republic show lower levels of trust in their municipalities, but would give
decentralization a chance. In Peru, Honduras and El Salvador, in spite of the fact that people trust
their municipalities more than the national government, they show more support for the process
of centralizing the state. In the case of Panama, Brazil and Mexico, there is practically no
109
distinction between municipal and national government trust, but people prefer that the national
government assume more municipal obligations and services. To conclude, the people of Chile,
Colombia, Haiti, Uruguay and Venezuela trust their national governments more and support state
centralization.
A factor to be considered in subsequent studies is the decentralization of the judicial

branch of government. All the literature reviewed for this study mentions three factors in
decentralizationadministrative, political and fiscalbut in all of them decentralization of
judiciary institutions has not been addressed. In the end, a complementary analysis of this
investigation would consist of determining the levels of balance of power in different levels of
government by way of formal modeling and game theory.
110
References
Agresti, Alan and Barabara Finlay. Statistical Methods for the Social Sciences. International Edition ed.
New Jersey: Prentice-Hall, Inc., 1997.
Barczak, Monica. "Representation by Consultation? The Rise of Direct Democracy in Latin America."
Latin American Politics and Society 43, no. 3 (2001): 24.
Bardhan, Pranab. "Decentralization of Governance and Development." Journal of Economic Perspectives
16, no. 4 (2002): 20.
Barr, Robert. "Parties, Legitimacy and the Motivations for Reform: Devolution and Concentration in Latin
America." (2001): 27. Publication data?
Bunce, Valerie. "Comparative Democratization: Big and Bounded Generalizations." Comparative
Political Studies 33, no. 6/7 (2000): 32.
Cameron, John. "Municipal Democratization in Rural Latin America: Methodological Insights from
Ecuador." Bulletin of Latin American Research 24, no. 3 (2001): 24.
Campbell, Tim. The Quiet Revolution: Decentralization and the Rise of Political Participation in Latin
American Cities. Pittsburgh: University of Pittsburgh Press, 2003.
Chang, Eric. "Electoral Incentives for Political Corruption under Open-List Proportional Representation."
The Journal of Politics 67, no. 3 (2005): 15.
Charles, Andrian and James Smith. Political Democracy, Trust and Social Justice. Boston: University
Press of New England, 2006.
Christopher Garman, Stephan Haggard, and and Eliza Willis. "Fiscal Decentralization: A Political Theory
with Latin American Cases." World Politics 53 (2001): 32.
CIA, Central Intelligence Agency. "The World Factbook." Washington, 2007.
Dahl, Robert. Polyarchy; Participation and Opposition. New Heaven: Yale University Press, 1971.
Daughters, Robert, Leslie Harper. "Fiscal and Political Decentralization Reforms." In The State of State
Reform in Latin America, edited by Eduardo Lora, 87 - 121. Washington: Stanford University
Press, 2007.
Epstein, David. "Democratic Transitions." American Journal of Political Science 50, no. 3 (2006): 19.
Escobar-Lemmon, Maria. "Political Support for Decentralization: An Analysis of the Colombian and
Venezuelan Legislatures." American Journal of Political Science 47, no. 4 (2003): 15.
Falleti, Tulia. "A Sequential Theory of Decentralization: Latin American Cases in Comparative
Perspective." American Political Science Review 99, no. 3 (2005): 20.
Freire, James Keese and Marco. "Decentralisation and NgoMunicipal Government Collaboration in
Ecuador." Development in Practice 16, no. 2 (2006): 14.
Garman, Christopher, Stephan Haggard, and and Eliza Willis. "Fiscal Decentralization: A Political Theory
with Latin American Cases." World Politics 53 (2001): 32.
Grote, Joachim vonBraun and Ulrike. "Does Decentralization Serve the Poor?" IMF (2000): 32.
Hetherington, Marc. Why Trust Matters? New Jersey: Princeton University Press, 2005.
Hiskey, Jonathan, and Mitchell Seligson. "Pitfalls of Power to the People: Decentralization, Local
Government Performance, and System Support in Bolivia." Studies in Comparative International
Development 37, no. 4 (2003): 25.
Joseph Tulchin, Andrew Selee. Decentralization and Democratic Governance in Latin America.
Washington: Woodrow Wilson International Center for Scholars, 2004.
Kenneth Newton, Pippa Norris. "Trust in Public Institutions: Faith, Culture or Performance?" (2000).
This looks like an incomplete entry
Lagos, Martha. "How People View Democracy: Between Stability and Crisis in Latin America." Journal
of Democracy 12, no. 1 (2001): 9.
Licha, Isabel. "Citizen Participation and Local Government in Latin America: Advances, Challenges and
Best Practices." IADB (2002): 8.
111
Lucero, Jose. "High Anxiety in the Andes: Crisis and Contention in Ecuador." Journal of Democracy 12,
no. 2 (2001): 15.
Manor, James. The Political Economy of Democratic Decentralization. Washington, 1999.
Nessim, Heli, Ana Mara Linares, Iveta Ganev, Diego Buchara, Julio Norori, Javier Aguilar, Silvia
Echeverra, Huscar Eguino. "Programa Multifase De Desarrollo Municipal Y Apoyo a La
Descentralizacin ", 15: BID, 2003.
Newton Kenneth and Pippa Norris. "Trust in Public Institutions: Faith, Culture or Performance?" Harvard
University (2000): 26.
ONeill, Kathleen. "Decentralization as an Electoral Strategy." Comparative Political Studies 36, no. 9
(2003): 24.
O'Neill, Kathleen. Decentralizing the State: Elections, Parties, and Local Power in the Andes. New York:
Perez, Maria Garcia-Guadilla and Carlos. "Democracy, Decentralization and Clientelism: New
Relationship and Old Pratices " Latin American Perspectives 29, no. 5 (2002): 21.
Prez-Lin, David Altman and Anbal. "Assessing the Quality of Democracy: Freedom, Competitiveness
and Participation in Eighteen Latin American Countries" Democratization 9, no. 2 (2002): 15.
format??
Peruzzotti, Catalina Smulovitz and Enrique. "Societal Accountability in Latin America." Journal of
Democracy 11, no. 4 (2000): 12.
Radcliffe, Sarah, Nina Laurie and Robert Andolina "Reterritorialised Space and Ethnic Political
Participation: Indigenous Municipalities in Ecuador." Space & Polity 6, no. 3 (2002): 17.
Roberts, Kenneth. "Social Inequalities without Class Cleavages in Latin Americas Neoliberal Era."
Studies in Comparative International Development 36, no. 4 (2002): 31.
Sadoulet, Alain de Janvry and Elisabeth. "Rural Poverty in Latin America: Determinants and Exit Paths."
Food Policy 25 (2000): 21.
Selee, Andrew. "Exploring the Link between Decentralization and Democratic Governance." In
Decentralization and Democratic Governance in Latin America, edited by Andrew Selee Joseph
Tulchin, 35. Washington, 2004.
Seligson, Mitchell. "The Impact of Corruption on Regime Legitimacy: A Comparative Study of Four
Latin American Countries." The Journal of Politics 64, no. 2 (2002): 26.
Seligson, Mitchell , Polibio Cordoba. Auditora De La Democracia Ecuador. Quito: Cedatos, 2006.
. Audiutora De La Democracia Ecuador. Quito: Cedatos, 2004.
Sorensen, Georg. Democracy and Democratization: Processes and Prospects in a Changing World
Westview Press, 1998.
Sukhtankar, Carol Graham and Sandip. "Does Economic Crisis Reduce Support for Markets and
Democracy in Latin America? Some Evidence from Surveys of Public Opinion and Well Being."
Journal for Latin American Studies (UK) (2004).
Treisman, Daniel. "Decentralization and Inflation: Commitment, Collective Action or Continuity." The
American Political Science Review 94, no. 4 (2000): 22.
. "Defining and Measuring Decentralization: A Global Perspective." (2002): 38. Publication data?
Zphyr, Dominique, Yves-Franois Pierre, and Abby Crdova. Culture politique de la dmocratie en
Hati: 2006. Edited by Latin American Public Opinion Project. Nashville, TN: Vanderbilt
University, 2007.
112
Appendix 1
LS3. Changing the subject, in general how satisfied are you with your life? Would you say that you are
..? (1) Very satisfied (2) Somewhat satisfied (3) Somewhat dissatisfied (4) Very dissatisfied (8) DK
IT1. Now, speaking of the people from here, would you say that people in this community are generally
very trustworthy, somewhat trustworthy, not very trustworthy or untrustworthy ..?
NP1. Have you attended a town meeting, city council meeting or other meeting convened by the mayor
in the past 12 months?
SGL1. Would you say that the services the municipality is providing are? [Read options]
(1) Very good (2) Good (3) Neither good nor poor (fair) (4) Poor (5) Very poor (8) Doesnt know
SGL2. How have they treated you or your neighbors when you have had dealings with the municipality?
Have they treated you very well, well, neither well nor badly, badly or very badly? (1) Very well (2)
Well (3) Neither well nor badly (4) Badly (5) Very badly (8) Doesnt know
VIC1. Have you been a victim of any type of crime in the past 12 months?
(1) Yes [continue] (2) No [go to AOJ8] (8) DK [go to AOJ8]
EXC2. Has a police official ask you for bribe during the past year?
EXC6. During the past year did any public official ask you for a bribe?
EXC11. During the past year did you have any official dealings in the municipality/local government?
If the answer is No mark 9
If it is Yes ask the following:
During the past year, to process any kind of document (like a license, for example), did you have to pay
any money above that required by law?
EXC13. Are you currently employed?
At your workplace, did anyone ask you for an inappropriate payment during the past year?
EXC14. During the past year, did you have any dealings with the courts?
If the answer is No note down 9
Did you have to pay a bribe at the courts during the last year?
EXC15. Did you use the public health services during the past year? If the answer is No mark 9
In order to receive attention in a hospital or a clinic during the past year, did you have to pay a bribe?
EXC16. Did you have a child in school during the past year?
Did you have to pay a bribe at school during the past year?
EXC17. Did anyone ask you for a bribe to avoid having the electricity cut off?
EXC18. Do you think that the way things are, sometimes paying a bribe is justified?
EXC19. Do you think that in our society paying bribes is justified because of poor quality public
services, or do you think it is not justified?
113
Appendix 2
Coefficients
Unstandardized Standardized
Coefficients Coefficients
Model t Sig.
B Std. Error Beta
1 (Constant) 6.540 1.423 4.596 .000
Sex (male) .439 .377 .007 1.165 .244
Age .367 .134 .018 2.746 .006
Education -.769 .287 -.020 -2.682 .007
Urban sector 1.162 .449 .018 2.589 .010
Wealth
-.187 .115 -.013 -1.616 .106
Interpersonal trust .077 .006 .078 12.031 .000

Life satisfaction .021 .007 .019 2.805 .005
Corruption victimization
-2.108 .250 -.056 -8.439 .000
Municipal service
perception .153 .019 .114 7.975 .000
Attend city hall
.424 .022 .282 19.475 .000
meetings
Delinquency
victimization -.019 .005 -.024 -3.742 .000
Ecuador 2.919 1.139 .024 2.564 .010

El Salvador 13.077 1.146 .108 11.415 .000
Guatemala 12.328 1.174 .097 10.501 .000
Honduras 8.723 1.170 .071 7.452 .000
Nicaragua 8.082 1.141 .065 7.085 .000
Peru 4.093 1.108 .034 3.693 .000
Dominican Republic 13.018 1.166 .106 11.160 .000
Chile 11.839 1.185 .097 9.991 .000
Costa Rica 7.004 1.209 .058 5.794 .000
Guyana 9.433 1.189 .069 7.932 .000
Jamaica 7.755 1.212 .054 6.399 .000
Panama 6.382 1.153 .050 5.537 .000
Mexico 15.672 1.167 .128 13.431 .000
Uruguay 11.564 1.199 .095 9.646 .000
Venezuela 2.874 1.255 .020 2.291 .022
a. Dependent variable: Trust in the Municipality
114
Appendix 3
N1. To what extent would you say the current administrationfights poverty.
N3. To what extent would you say the current administration promotes and protects democratic
principles.
N9. To what extent would you say the current administrationfights government corruption.
N10. To what extent would you say the current administration protects human rights.
N11. To what extent would you say the current administration improves the security of our citizens.
N12. To what extent would you say the current administration fights unemployment.
Now we want to know how much information about politics and the country is transmitted to the people...
GI1. What is the name of the current president of the United States? [Dont read, George Bush]
(1) Correct (2) Incorrect (8) Do not Know (9) No Answer
GI2. What is the name of the President of Congress in country? [Dont read, insert name]
GI3. How many provinces does dcountry have? [Dont read, insert number of provinces]
(1) Correct (2) Incorrecto (8) Do not Know (9) No Answer
NICARAGUA AND PANAMA ACCEPT WITH OR WITHOUT COMARCAS
GI4. How long is the presidential/prime minister term in country? [Dont read, insert number of years]
((1) Correct (2) Incorrect (8) Do not Know (9) No Answer
GI5. What is the name of the president of Brazil? [Dont read, Luiz Incio Lula da Silva, also accept
Lula]
SGL1. Would you say that the services the municipality is providing are? [Read options]
(1) Very good (2) Good (3) Neither good nor poor (fair) (4) Poor (5) Very poor (8) Doesnt know
SGL2. How have they treated you or your neighbors when you have had dealings with the municipality?
Have they treated you very well, well, neither well nor badly, badly or very badly? (1) Very well (2) Well (3)
Neither well nor badly (4) Badly (5) Very badly (8) Doesnt know
VIC1. Have you been a victim of any type of crime in the past 12 months?
(1) Yes [continue] (2) No [go to AOJ8] (8) DK [go to AOJ8]
EXC2. Has a police official ask you for bribe during the past year?
EXC11. During the past year did you have any official dealings in the municipality/local government?
During the past year, to process any kind of document (like a license, for example), did you have to pay any
money above that required by law?
EXC14. During the past year, did you have any dealings with the courts?
If the answer is No note down 9
Did you have to pay a bribe to the courts during the last year?
115
EXC15. Did you use the public health services during the past year? If the answer is No mark 9
EXC17. Did anyone ask you for a bribe to avoid having the electricity cut off?
EXC18. Do you think that the way things are, sometimes paying a bribe is justified?
EXC19. Do you think that in our society paying bribes is justified because of poor quailty of public services,
or do you think it is not justified?
Now, moving on to a different topic, sometimes the people and communities have problems that they cannot
solve by themselves, and so in order to solve them they request help from a government official or agency.
In order to solve your problems have you ever requested help or cooperation from...?
CP2. A member of congress/parliament
CP4A. Some local public official (e.g, a mayor, municipalcouncilperson, provincial official)
B14. To what extent do you trust the national government?
B32. To what extent do you trust the Mayors office of your municipality?
116
Appendix 4
Coefficients
B S.E. Wald df Sig. Exp(B)

Step Male .085 .031 7.634 1 .006 1.089
1(a) Age .057 .011 27.439 1 .000 1.058
Urban sector .193 .036 27.923 1 .000 1.213
Education .172 .025 47.993 1 .000 1.187
Wealth .006 .010 .377 1 .539 1.006
Government
-.003 .001 17.936 1 .000 .997
efficacy
Political
.004 .001 34.513 1 .000 1.004
Knowledge
Municipal
service .012 .001 185.704 1 .000 1.012
perception
Trust in the
.006 .001 116.514 1 .000 1.006
Municipality
Trust in the
Nacional -.001 .001 1.837 1 .175 .999
Government
Interest in
-.109 .017 40.742 1 .000 .897
Politics
Help request in
the mayors .001 .000 2.378 1 .123 1.001
office
Help request in a
.001 .001 3.664 1 .056 1.001
Ministry
Delinquency
.000 .000 .011 1 .917 1.000
victimization
Corruption
.014 .020 .481 1 .488 1.014
victimization
Ecuador 1.314 .099 177.803 1 .000 3.719
El Salvador .912 .096 89.407 1 .000 2.489
Guatemala 1.577 .101 245.647 1 .000 4.842
Honduras .685 .097 49.583 1 .000 1.983
Nicaragua 1.810 .100 327.056 1 .000 6.111
Peru 1.048 .097 117.442 1 .000 2.853
Dominican R. 1.543 .098 246.783 1 .000 4.680
Chile 1.505 .098 234.202 1 .000 4.503
CRica 1.221 .100 147.967 1 .000 3.392
Guyana 1.550 .105 216.838 1 .000 4.713
Jamaica 2.023 .109 345.640 1 .000 7.563
Haiti 1.022 .103 98.012 1 .000 2.780
Mexico 1.145 .102 126.660 1 .000 3.143
Uruguay .894 .101 79.082 1 .000 2.445
Brazil .753 .101 55.351 1 .000 2.124
Venezuela .961 .104 85.007 1 .000 2.615
Constant -2.695 .148 330.785 1 .000 .068
b. Dependent Variable: Decentralization versus Centralization.
117
VI. Is the Vote for the left a Risk or Opportunity for Democracy
in Latin America?
Rosario Queirolo
Abstract
What is the impact that market-oriented economic reforms have had on the vote for leftist parties
in Latin America? Are Latin Americans voting for the left based on their ideological stances or
because parties on the left merely benefit from voters discontent towards traditional parties?
Latin Americans can vote left because they want more state intervention in the economy, a more
egalitarian economic distribution, or more investment in social policies. Alternatively, it is
possible to argue that voters are not policy oriented, they only care about outcomes, and they are
voting left because the neoliberal model failed to deliver sustainable economic development and
to overcome the endemic problem of unemployment. My central argument is that the recent
increase of leftist parties in Latin America comes about as a result of voters punishing political
parties that were unable to improve the economic well-being of their electorates. In addition, this
article compares the attitudes towards democracy of those Latin Americans identified as leftists
with those that declared themselves as rightists to explore if the movement to the left represents a
risk or opportunity for democracy. The evidence from AmericasBarometer 2006-07 indicates that
leftists are not significantly less democratic than rightists.
Is the vote for the left a risk for democracy in Latin America? Since the final years of the
Twentieth Century, many Latin American countries have elected governments that identified
themselves with the ideological left. In 1999, Hugo Chvez, a former coup plotter, was elected
President of Venezuela after campaigning against the Washington consensus model, and
promising to upend the old social order and improve the lives of the poor. Brazil also veered
toward the left with the victory of the Workers Party (Partido dos Trabalhadores, PT) in the 2002
general elections. In Argentina, a left-wing political faction of the Peronist Party headed by
Nstor Kirchner won the 2003 election; while in neighboring Uruguay, the Broad Front (Frente
Amplio) a left-leaning coalition party which has steadily increased its electoral participation since
it was founded in 1971, finally gained the presidency in 2004. Chile has been governed by a
center-left coalition since its return to democracy; the chair of the government has alternated
between social democrats and socialists, and in the 2005 election a female socialist candidate
became President. Also in 2005, Bolivians decided to grant Evo Morales, the presidential
candidate of Movimiento al Socialismo (MAS), and an important leader of the coca producers
union, the chance to govern one of the poorest countries in Latin America. Manuel Lpez
Obrador, the presidential candidate for the Partido Revolucionario Democrtico (PRD) in
Mxico, lost the presidential election held in July 2006 by less than 1% of the votes in a very
controversial and disputed election. More recently, at the end of 2006, Nicaragua and Ecuador
Professor of Political Science at the Universidad de Montevideo.
119
have chosen leftist political parties to be in charge of the government. Daniel Ortega, former
president of Nicaragua from 1985 to 1990, and leader of the Frente Sandinista de Liberacin
Nacional (FSLN) was reelected as president in November 2006. In Ecuador, Rafael Correa won
the presidency in the second round of the election with the support of leftists political parties and
indigenous movements.
The movement of Latin America to the left has led journalists, political analysts and
political scientists to look for explanations. The most widespread of these suggests that Latin
Americans vote for political parties on the left is a backlash against the neoliberal model
implemented in the region mainly during the 1990s. A deeper concern is the implications of this
shift to the left for the stability of democracy under these new leftists governments. This chapter
attempts to explore both of these issues using AmericasBarometer 2006-2007 public opinion data
set.
To anticipate, the research comes to three main conclusions. First, Latin Americans are
not voting leftist parties because they are against neoliberal policies. The current shift to the left
in Latin America is far more a result of popular discontent with the voters economic situation
than anything else. Second, the electoral prospect that leftist parties have by capitalizing on social
discontent depend heavily upon the number of untainted opposition parties available in the
political system. In countries like Brazil and Uruguay where leftist parties embody the only
untainted opposition, it was easier for them to capitalize on popular discontent than in Mexico,
where a party on the right also represented an untainted opposition. Finally, rather than a threat
to the stability of democracy the findings of the research show that the rise of the left is actually
good news for the future of democracy in the region because Latin Americans have been able to
demonstrate to themselves, at least, that governments can now incorporate the formerly excluded
opposition left-oriented parties into the political game. Democracy is at risk when, as was the
case in Mexico for over 50 years, opposition parties are always on the losers side. Throughout
the region, for many years, several leftist political parties were consistent losers in the electoral
game, leading some of them to dismiss electoral democracy as a valid means to achieve power.
Therefore, the arrival of left-leaning parties to the government of several Latin American
countries, rather than being a cause of concern, should be considered an indicator of a healthy
democracy and a mechanism to strengthen democratic support among citizens. Latin Americans
are proving capable of making their political leaders accountable, removing them from office
when they do not accomplish what was expected of them, and changing those in charge of the
government by voting for untainted parties. The recent shift towards the left in Latin America
has helped to intensify and strengthen democracy in the region by incorporating losers into the
political game.
The first section of this chapter discusses the argument that the vote for leftist parties in
Latin America is a backlash against the neoliberal model implemented in the region. This
argument is tested for three country cases in which leftist parties have been increasing their
electoral support during the last decade: Brazil, Mexico and Uruguay. The second section
describes each of these country cases and states the theories and hypotheses used to evaluate the
argument. In the third section, I present the models and results of the statistical tests which show
that Latin-Americans vote for left-of-center parties is not a rejection of market-oriented policies
but a reaction to the state of the economy. The second argument tested is that the vote for the left
120
represents a risk for the embryonic Latin American democracies. Rather than weaken democratic
prospects, the success of several leftist parties in the region makes democracy stronger by
channeling social discontent through institutional mechanisms. The fourth section explains this
argument and states the hypotheses used to assess its validity. The fifth section shows the
statistical results of these tests performed using the Americas Barometer 2006-2007 dataset. To
conclude, I discuss the implications for democracy and political accountability in the region.
First Argument: Voting Left as a Backlash Against the Washington Consensus
During 1980s and 1990s, the neoliberal model based on the so-called Washington
Consensus, was implemented to various degrees in Latin American countries. Market oriented
economic reforms were initiated in the 1980s, or even earlier in Chile, Uruguay, Argentina, and
Colombia (Morley, Machado and Pettinato 1999).1 In this chapter the terms Washington
Consensus, market-oriented economic reforms, structural reforms, neoliberal model, or
orthodox policies are used interchangeably, and it is assumed that all of these terms refer to the
same set of policies described by Williamson as the so-called Washington Consensus. To
remind the reader, the set of policy reforms grouped as Washington Consensus can be
summarized as involving fiscal discipline, public expenditure restrictions, tax reform, interest rate
liberalization, a competitive exchange rate, trade liberalization, liberalization of restrictions on
foreign direct investment, privatization, deregulation, and reinforcement of property rights.
Many scholars have undertaken the task of measuring the success or failure of the
Washington Consensus (Dutch 2003; Escaith and Morley 2001; Huber and Solt 2004; Kuczynski
and Williamson 2003; Lora and Panizza 2002; Lora, Panizza and Quispe-Agnoli 2004; Stallings
and Peres 2000), and many others have analyzed the impact of particular policy reforms (Lora
and Barrera 1997). Regardless of the differences between those studies, they agree that after two
decades of reforms the expected results of economic growth was far from universally achieved.
While all agree that fiscal responsibility has spread throughout the region, and hyper inflation
seems to be thing of the past, sustainable economic growth and, for our purposes more
importantly, improvement of social indicators have been illusive. The Chilean case is an
important exception to the overall uneven pattern, but it is difficult to point to other cases in
which growth and social indicators have improved steadily.
Regardless of the objective outcomes of the neoliberal model, Latin Americans

disenchantment with it can be seen everywhere. Many of the strongest supporters of the model
have recognized that the outcome was not the one that they were hoping for. International
organizations, which strongly supported the neoliberal model such as the International
Monetary Fund, the World Bank and the Inter-American Development Bank, have acknowledged
that the reforms did not produce the expected results, and they now suggest four different types of
follow-up reforms to overcome this failure: crisis proofing, completing first-generation reforms,
advancing second-generation reforms, and improving equality (IDEA 2004). From this
1
Morley et al. (1999) point out that most of the rise in the trade and financial reform indexes during the 1970s are
due to the policies implemented in Chile, Uruguay, Argentina, and Colombia.
121
perspective, it was not the reforms that were at fault, but the failure to implement them fully and
effectively that these follow-up reforms are attempting to address. Many scholars who initially
supported the Washington Consensus as the way to achieve development later moved away
from this idea, and became its critics: Jeffrey Sachs (2005), Joseph Stiglitz (2002) and Dani
Rodrik (2001) are leaving examples of this shift. Some have argued that the international
community has shown signs of reform fatigue (Edwards 1997).2 The lack of public support for
the Washington Consensus can also be seen among the general public; there is a widespread loss
of confidence in the benefits of pro-market reforms among opinion leaders, and a less proactive
stance toward reforms is the current mainstream among Latin Americas policymakers (Lora,
Panizza and Quispe-Agnoli 2004; Panizza and Yaez 2005).
In the view of many political analysts, the current increase in the vote for the left in Latin
America is a consequence of reform fatigue. Simply stated, this argument says that because
voters are tired of market-oriented economic reforms and their consequences, they are voting in
favor of parties that allow more state intervention in the economy. The Economist magazine
states this argument as follows: Rightly or wrongly, voters blamed the slowdown on the free-
market reforms known as the Washington Consensus. As happens in democracies, they started to
vote for the opposition- which tended to be on the left. (The Economist 2006).
However, this is not the only answer. Others have pointed out that, behind this shift to the
left, there lies primarily a need for a change. Popular discontent at traditional parties unable to
solve problems of poverty, corruption and inequality led Latin Americans to vote for political
parties perceived as being more likely to deliver a better standard of living. To put it simply,
according to this view, Latin Americas shift to the left is rooted less in ideological stances than
in a desire to punish incumbents for poor economic performance.
This chapter disentangles what is true in each of these arguments. What is the impact that
market-oriented economic reforms have had on the vote for leftist parties in Latin America? Are
Latin Americans voting for the left depending on their ideological stances or because parties on
the left merely benefit from voters discontent towards traditional parties? Latin Americans can
vote left because they want more state intervention in the economy, a more egalitarian economic
distribution, or more investment in social policies. After a decade of neoliberal economic
reforms, they may be claiming that its time for a change (Schlesinger 1986), and consequently,
may behave in a policy-oriented way. Alternatively, it is possible to argue that voters are not
policy oriented, they only care about outcomes, and they are voting left because the neoliberal
model failed to deliver sustainable economic development and to overcome the endemic problem
of unemployment. These two explanations are not incompatible, both can be true. Latin
Americans may be voting left because they do not want more market-friendly economic policies,
and also because they are punishing incumbent parties for poor economic performance.
My central argument is that the recent increase of leftist parties in Latin America comes
about as a result of voters punishing political parties that were unable to improve the economic
2
Reform fatigue is a concept coined by Sebastian Edwards (1997) that encompasses citizens tiredness with the
sacrifices required by economic reforms in their respective countries.
122
well-being of their electorates. Most Latin Americans have faced economic hardship during
successive governments under a variety of political parties, and recent research demonstrates that
voters have long-term economic memories (Benton 2005) and punish not only the incumbent
party for the material suffering; they also rebuke parties that governed before the incumbent came
to power. Left-of-center parties took advantage of this popular discontent and capitalized on
social and economic dissatisfaction when they were outside the governing coalitions and
remained in the opposition. As a result, by voting left-oriented parties, Latin Americans seem to
be looking for credible political alternatives to the status quo rather than becoming anti-market in
their policy positions.
Theories, Hypotheses and Country Cases
Latin American voting behavior is usually understood as being highly volatile and
unpredictable due to the lack of strong party and ideological identifications. Latin Americans
seem mainly to base their vote choice on short-term factors such as economic conditions (Cantn
and Jorrat 2002; Roberts and Wibbels 1999) and candidate image (Echegaray 2005; Weyland
2003). It is within this context that the recent victories of leftist parties have become puzzling. If
ideology and party identification are not relevant voting clues in Latin America (Echegaray
2005), why are voters choosing parties identified with the ideological left? Is the vote for leftist
parties another example of economic voting theory according to which voters punish the
incumbent party for poor economic results? Are electorates in Latin America mainly choosing
leftist parties because their candidates are, on average more appealing than are the candidates
from parties of the center and right? Or, alternatively, are Latin-Americans becoming more
ideological and policy-oriented by voting for the left because ideology does indeed matter and
voters are rejecting the neoliberal paradigm?
Taking into account the research on voting behavior done on Western Europe and the
United States, and building on the results of previous studies about Latin American voting, I will
test two theories to explain the rise of the left in Latin America: economic voting theory, and the
cleavages created by political processes (Przeworski and Sprague 1986; Torcal and Mainwaring
2003). In particular, the individual level analysis aims to understand what the role is of economic
evaluations (economic voting theory), and ideology and policy issues (cleavages created by
political processes) in the recent rise of the left in Latin America. At the same time, I will be
testing if Latin Americans are policy-oriented (ideology and policy issues are significant
determinants of the vote), outcome-oriented (economic evaluations are the significant predictors
of the vote), or both. The following sub-sections briefly describe each of the said theories,
summarize the major research done in Latin America using each of them, and go over the main
hypotheses and variables by which the theories are going to be tested.
Economic voting theory
The literature on voting behavior in Latin America is dominated by the economic voting
explanation. Economic voting theory states that if the economy is doing fine, voters will reelect
123
the incumbent party; while in bad times, citizens will punish the incumbent at the ballot box. The
theory has taken four major forms: pocketbook vote, sociotropic vote (Kinder and Kiewiet 1981),
retrospective vote, and prospective vote (MacKuen, Erikson and Stimson 1992). These
distinctions lead to four possible combinations in which citizens can appraise the economic
situation: evaluating how good or bad the economic situation of the country has been during the
past (retrospective sociotropic), taking into account voters expectations of how the countrys
economic situation is going to be in the future (prospective sociotropic), thinking on how good or
bad their familys economic situation has been in the recent past (retrospective pocketbook), or
considering their expectations for their familys economic future (prospective pocketbook) .
Economic voting theory has noticeably proved its predictive power in the stable economic
and political contexts of the United States and Western Europe (Fiorina 1981; Kinder and
Kiewiet 1981; Lewis-Beck 1986; Lewis-Beck 1988; Lewis-Beck and Belluci 1982; MacKuen,
Erikson and Stimson 1992; Nadeau and Lewis-Beck 2001). And there is a consensus regarding
the idea that Americans and Europeans respond to changes in general economic conditions
much more than to changes in the circumstances of personal economic life (Kinder 1998).
In Latin America, scholars have tested the relationship between economic downturns and
voting for incumbent parties in single-country case studies (Cantn and Jorrat 2002; Domnguez
and McCann 1995; Mora y Araujo and Smith 1984; Remmer 2003; Roberts and Arce 1998;
Seligson and Gmez 1989; Weyland 1998), and through comparative studies (Echegaray 2005;
Remmer 1991; Roberts and Wibbels 1999; Remmer 1993), but the evidence is far from
conclusive. Economic evaluations matter for Latin-Americans depending on the election. For
example, Weyland found that Venezuelans were Pocketbook voters from 1989 to 1993 (Weyland
1998), but Sociotropic voters when they elected Hugo Chvez in 1998 (Weyland 2003). Cantn
and Jorrat (2002) and Echegaray (2005) also find that the impact of the economy on Latin
Americans vote choice varies across countries and elections. Despite these distinctions, scholars
confirm that voters in Latin America tend to treat elections as plebiscites on the economic
performance and capabilities of the government.
If economic factors are important determinants of the fortune of incumbents, are there any
specific economic conditions that favor leftist parties in comparison with centrist or rightist
parties? Following the economic voting explanation, I expect that voters who evaluate negatively
the economic situation will punish the incumbent. In countries where the incumbent is a leftist
party, citizens will reward or punish it depending on the economic performance. But in countries
where leftist parties were never in charge of the government and represent a credible or
untainted opposition, electorates which are economically dissatisfied with the economy will
cast their vote in favor of them.
H1: The more negatively a voter evaluates the national economic situation, the greater
the probability he or she will vote for the opposition. In particular, voters who are
discontented will reward leftist parties when they were not in charge of the government.
To put it simply, if a voter has a negative economic evaluation (x1) and leftist parties
represent a credible or untainted opposition (x2), he or she will vote for the left (y).
124
The variables to test the economic voting theory are four: retrospective sociotropic vote,
retrospective pocket-book vote, prospective sociotropic vote and retrospective sociotropic vote.
Sociotropic vote measures the evaluation of the countrys economic situation; the higher the
value, the worse the evaluation. Pocketbook vote measures the evaluation of the familys
economic situation; the higher the value of the variable, the worse the familys economic
assessment is. Prospective measures the expectations regarding the economic future, while
retrospective measures the evaluation of the country economic situation in comparison with the
past.
Cleavages Created by Political Process
In determining which theory and voting clues best explain the vote for leftist parties in
Latin America, it is essential to include ideology. Ideology is regarded as one of the most
influential voting clues. Electorates use the overarching continuum between left and right, or
from liberal to conservative, as a shortcut to processing political information and making their
electoral decisions. Since Converse (1964) there has been a great deal of debate about how
readily voters rely on ideology when voting, and to what extent citizens organize their political
opinions around the ideological dimension. The same doubts are cast regarding the importance of
ideology to predict Latin Americans voting behavior. Echegaray (2005) considers that
ideological clues are an irrelevant source of guidance for Latin American voters, but he does not
empirically test this contention. Differing from Echegarays position, this study will test the
impact of ideological clues on the vote for leftist parties. Two main reasons make the inclusion of
ideology reasonable.
First of all, around seven out of ten Latin Americans were able to place themselves in the
ideological dimension according to the AmericasBarometer for 2006-2007. This percentage
varies depending on the country; left and right ideological labels mean more to Chileans and
Uruguayans than to Argentineans. But, as a first appraisal, ideological thinking is part of most
Latin Americans political behavior. Second, there is empirical research pointing to ideology as a
relevant voting clue for Latin Americans (Cameron 1994; Torcal and Mainwaring 2003). Torcal
and Mainwaring (2003) point out: class emerges as a major cleavage in party systems to the
extent that parties of the left emphasize class issues, and they called this phenomenon the
cleavage created by political processes (Przeworski and Sprague 1986; Torcal and Mainwaring
2003).3 This theoretical approach to cleavage formation pays attention to how cleavages are
created by political elites and political factors. The left/right ideological division can also be
considered a cleavage created by political process. In other words, politicians can activate this
cleavage as a way to get votes. All this suggests that it is appropriate to test for ideological clues:
H2: Ideological self-placement is likely to determine the vote for the left irrespective of
social and structural determinants.
3
Torcal and Mainwaring (2003) test the existence of these political cleavages in the Chilean case with three cultural-
ideological divisions that can be used by political leaders to articulate conflict: the authoritarian/democratic cultural
division, the perception of social inequality, and religious differences.
125
An alternative way to test the ideological cleavage is to analyze if policy positions are
determinants of voting behavior. Voting for the left is usually associated with support for
government involvement and regulation of the economy, income redistribution, and an increase
in social spending (Fuchs and Klingermann 1990; Inglehart and Klingermann 1976; Kitschelt and
Hellemans 1990). In addition, and due to the difficulty in obtaining survey data that deals with
citizens perceptions and opinions towards market-oriented economic reforms, the analysis of
policy preferences is the best way to approach this issue. Consequently, I hypothesize that:
H3: Those Latin Americans who support government involvement and regulation of the
economy, income redistribution and an increase in social spending will be more likely to
vote for leftist parties, while those who are against these policy issues will be more likely
to vote for rightist parties.
Ideology is measured by the ideological self-placement of the respondent in a dimension

that ranges from 0 meaning Left, to 10 meaning Right. Different policy issues are also used
as independent variables to test the ideological cleavage: support for privatization, support for
agrarian reform, support for nationalization, opinion towards redistribution, opinion about state
regulations and state interventionism.
The Cases
My argument that the recent increase of leftist parties in Latin America comes about as a
result of voters punishing political parties that were unable to improve the economic well-being
of their electorates and rewarding untainted political parties is tested in three country cases:
Brazil, Mexico and Uruguay. I argue that leftist parties can capitalize social discontent when: 1)
they represent a credible or untainted opposition, and even more so when 2) they are the only
untainted opposition in the political system. A credible or untainted political party is a party
that was never in charge of the government and cannot be held responsible for the countrys
welfare.
Brazil, Mexico and Uruguay are all cases in which leftist parties have increased their
share of the vote since 1980s, but the electoral trajectories followed by left-of-center parties in
each country differ. In Brazil, leftist parties gained access to the government in 2002. Before that,
in 1994, the Partido Social Democracia Brasileira (PSDB), a social democrat party, carried
Fernando Henrique Cardoso to the presidency. However, when the PSDB was elected to Brazils
national government, it had already moved to the right of the ideological scale (Power
2001/2002). Therefore, the first time that a left-of-center party gained access to Brazils national
government after the return to democracy was in 2002 through Luis Incio Lula da Silva, the
long-time leader of the Partido dos Trabalhadores (PT). After the experience of the PSDB
government, the most credible opposition was embodied by the PT.
In Uruguay, leftist parties have progressively increased their electoral participation since
the return to democracy in 1984, and after twenty years of democracy, in 2004 a left-leaning
coalition called the Encuentro Progresista-Frente Amplio (EP-FA) won the presidency. The
126
search for new alternatives has led Uruguayans to vote for the Frente Amplio, a left-leaning
coalition party which represents the only credible or untainted opposition after a long
succession of Partido Colorado and Partido Nacional governments.
Mexico represents a different example for the same phenomenon. Leftist parties, in
particular the Partido Revolucionario Democrtico (PRD), have increased their share of the vote
during the 1990s, and by doing so, have helped to raise competitiveness in the Mexican electoral
arena. The PRD received almost a fifth of the votes cast in the 1994 and 2000 presidential
elections, and in 2006 it lost the presidency by fewer than 500,000 votes in a highly controversial
vote count. However, the electorates search for something new ended up with their favoring the
two credible and untainted opposition parties: PAN, a center-right political party and the PRD, a
leftist party.
All these leftist parties, PT, PRD, and EP-FA, are examples of professional parties: they
care about party building, they have relatively strong party organizations, and they mobilize
political support in addition to social support. In that sense, they are more similar to
Concertacin in Chile than to Movimiento al Socialismo (MAS) in Bolivia or Hugo Chvezs
party, the Movimiento Quinta Repblica in Venezuela. They are usually categorized as the
institutional left in Latin America, contrary to the populist left represented mainly by the
Movimiento Quinta Repblica. Regardless of these commonalities, there are several differences
between these countries that make the case selection relevant. I will only refer to those
characteristics that are pertinent for the purpose of this research: differences in their party and
political systems, and differences in the level of economic reforms.
The differences between Brazilian, Mexican and Uruguayan party and political systems
are large. Brazil is usually defined as a case of party underdevelopment and weakly-established
political institutions (Ames 2001; Mainwaring 1999; Mainwaring and Scully 1995). Its
multiparty system has been described as highly fragmented, electoral volatility is comparatively
high, more than one-third of sitting legislators change parties during a term, and individualism,
clientelism and personalism rather than programmatic appeals dominate electoral campaigns
(Samuels 2006). For a long time, Mexico was characterized as a weakly-institutionalized political
system (Mainwaring and Scully 1995) with single-party dominance in the shape of the long-
ruling Partido Institucional Revolucionario (PRI). The PRI was in charge of the national
government from 1929 to 2000, and opposition parties were unable to win a majority in the lower
chamber of congress until 1997 when the single, dominant-party system was broken in favor of a
multiparty system. It was not until 2000 that the Partido de Accin Nacional (PAN), a right-
leaning party, ousted the PRI from the presidency. Uruguay has had a very stable party system
(Mainwaring 1995), with three major political parties, Partido Colorado (PC), Partido Nacional
(PN) and Frente Amplio (FA), and one minor party, Partido Independiente (PI). It was with the
emergence of the Frente Amplio in 1971 that the party system experienced a major change
evolving from a two-party system to a multiparty system (Gillespie and Gonzlez 1989;
Gonzlez 1991).
To put it simply, Brazil, Mexico and Uruguay are dissimilar in their levels of party system
institutionalization and numbers of political parties. Mainwaring and Scully (1995) classify
Brazil as an inchoate party system, Mexico as a hegemonic party system, and Uruguay as an
127
institutionalized one. Several things changed by the end of the 1990s - one is that Mexico can no
longer be considered a hegemonic party system. In terms of the number of parties, Mexico and
Uruguay have experienced important transformations by becoming multiparty systems and
raising their level of party competition. Recent research shows that the number of parties affects
the way in which voters hold governments accountable; multiparty systems strengthen voters
ability to punish several parties at a time, and therefore, popular discontent may be lower in
countries with more permissive electoral rules that allow small parties to gain congressional
representation (Benton 2005).
Market-oriented economic reforms were also implemented very differently in Brazil,

Mexico and Uruguay. Brazil and Mexico are classified as slow reformers: they started reforms
later and adopted less structural reforms; while Uruguay is considered a gradual reformer:
reforms were adopted earlier, but they were milder and carried out in a gradual way (Lora
1997/2001). The differences in the reforms pursued in Brazil, Mexico and Uruguay also depend
on the area being reformed. Brazil presents some of the highest privatization reform and labor
reform indexes. On the other hand, Mexico ranks low on their tax reform and labor reform
indexes, but high on the financial reform index. Finally Uruguay has one of the lowest levels of
privatization in the region but one of the highest indexes of trade reform (Lora 1997/2001).
The Model and Results
In order to test if policy preferences or ideological identifications are more important than
economic assessments when Latin Americans vote leftist parties, I use survey data collected in
the three country cases at the time of their pivotal elections. The Uruguayan data comes from a
pre-election survey carried out by CIFRA, Gonzlez, Raga y Asociados before the 2004 election,
the election in which the left got access to the government.4 The survey is national and includes
1,500 respondents. The comparison between the percentage intending to vote for leftist parties in
this survey with the proportion that actually voted for the left when the elections were held,
strengthen the validity of the analysis: the election result was 54% and the survey predicted a
60%.
To test the hypotheses in the 2002 Brazilian presidential election, when PT won for the first
time the presidency, I use data from Brazils 2002 National Election Study (BNES), a national
post-election voter behavior survey which includes 2,513 respondents5. Because this survey does
not include questions about economic evaluations, I also test the model with a 1998 pre-election
national survey carried out by Datafolha which includes 4,380 cases and the data collection
occurred during July. As is the case for Uruguay, Brazilian survey data also fits very well the
proportion intending to vote left with the proportion that actually voted left: in 2002, 68% of the
respondents said that had voted left and the actual percentage was 77%6.
4
I would like to thank the directors of CIFRA, Luis E. Gonzlez and Adriana Raga for giving me access to this data,
and generously allowed me to include some specific questions in the 2004 survey.
5
I want to thank Rachel Meneguello and Simone Aranha from the Center for Studies on Public Opinion (CESOP) at
the University of Campinas (UNICAMP) in Brazil for giving me access to Datafolha and BNES data.
6
This result is counterintuitive because post-election surveys usually over represent the winner.
128
The vote for leftist parties in the 2000 Mexican presidential election is analyzed through the
Post-Electoral Cross-Section survey carried out as part of the Mexico 2000 Panel Study7. This
survey includes 1,199 cases collected from July 6 to July 9 at respondents homes. As for Brazil
and Uruguay, the proportions intending to vote left according to the Mexican data correspond
very closely with proportions actually voting left when the elections were held, ensuring the
validity of the analysis. For the 2000 Mexican post-electoral survey, it is only possible to separate
the vote for PRD (not for others leftist parties), and the comparison between the survey and
election proportions is the following: 15% to 19% respectively. In all the surveys, data was
collected by personal, door-to-door interviews in the respondents homes8.
The dependent variable is a dummy variable that measures the intention to vote for a left-
of-center party, value 1 means that the person intended to vote for the left, while 0 represents the
vote intention for the remaining political parties. The following political parties were classified as
left-of-center in the presidential elections analyzed: in Uruguay, the Frente Amplio and Nuevo
Espacio/Partido Independiente; in 1998 Brazil, the PT, PPS and PSTU; in 2002 Brazil, the PT,
PSB, and PSTU; and in Mexico only PRD because it was not possible to separate the vote for
other leftist parties that have been put together under the other category in the dataset.
I have used two independent variables, sociotropic vote and pocketbook vote, to test the
economic voting theory. Sociotropic vote measures the evaluation of the countrys economic
situation; thus, the higher the value, the worse the evaluation. Pocketbook vote measures the
evaluation of the familys economic situation. The Retrospective Sociotropic, Prospective
Sociotropic, Retrospective Pocketbook and Prospective Pocketbook variables measure citizens
economic assessments of the country and their own situation in comparison with the past and
economic expectations for the future. The Sociotropic vote and Pocketbook vote in 1998
Brazilian survey, measure respondents evaluation of the Plano Real for the country and for
voters own life. Higher values correspond to negative evaluations.
Ideology is one of the independent variables I used to test the political cleavage; I measured
it by situating the ideological self-placement of the respondent in a dimension that ranges from
1 meaning left to 10 meaning right. Ideology is measured in two ways. The other way to
measure an interviewees ideology and test the political cleavage is through a series of questions
asking citizens opinions toward a series of policy issues: state interventionism, redistribution,
state regulations of private firms, agrarian reform, nationalization, and privatizations. Higher
values in each of these policies correspond with liberal positions, which I expect to be negative
correlated with the vote for leftist parties.
Other variables are included in the model as control variables: age, education, family
income, household level, and urban voter (residence). Age, education and family income have a
7
Participants in the 2000 Mexico Panel Study included (in alphabetical order): Miguel Basaez, Roderic Camp,
Wayne Cornelius, Jorge Domnguez, Federico Estvez, Joseph Klesner, Chappell Lawson (Principal Investigator),
Beatriz Magaloni, James McCann, Alejandro Moreno, Pablo Pars, and Alejandro Poir. Funding for the study was
provided by the National Science Foundation (SES-9905703) and Reforma newspaper.
8
The sample size of each survey is reduced after deleting all the missing values. The final N of is reported in the
tables that present the regression results.
129
straightforward interpretation; low values denote young people, low education, and low income.
Household level is measured in two ways. First, as an ordinal variable that classifies the
interviewees in categories based on an indicator of their household, higher values correspond to
higher socioeconomic level; and second as a houseware index in which higher values also
correspond with high socioeconomic level. Urban Voter is a dummy variable representing the
region in which the respondent lives; it takes the value of 1 when the person lives in an urban
area and 0 when he/she lives in a rural area or in other smaller cities and towns.
Finally, I also include partisanship as control variable using four dummy variables; each
dummy represents one category of partisanship: party identification with left-of-center parties,
party identification with parties at the center, party identification with right-of-center parties and
those that lack partisanship. Each category is entered into the model as a dummy variable that
takes the value of 1 when the person belongs to it and 0 when he/she does not. Those that have no
partisanship are the base category in the regression.
Results
The first thing to notice is that economic voting explains Latin Americans vote for leftist
parties in Brazil and Uruguay, but not in Mexico. Regressions results in Table VI-1 indicate that
Brazilians and Uruguayans who negatively evaluated the economic situation, tended to vote more
for left-of-center parties than for center or rightist ones. Citizens who were discontented with the
results that the economy had on their own lives, or in the countrys well-being, voted for left-of-
center parties, while those that made a positive evaluation reelected the incumbent government or
voted for other non-leftist party. The positive signs on the sociotropic and pocketbook
coefficients in Table 1 indicate that the worse the economic evaluation, the higher the probability
to vote for the left. This evidence supports hypothesis 1, which states that the more negatively a
voter evaluates the national economic situation, the greater the probability he or she will vote for
the opposition. In particular, voters who are discontented will reward leftist parties when they
were not in charge of the government. In Brazil and Uruguay, leftist parties were not responsible
for the government until 2002 and 2004, respectively.
However, Mexico presents a different case. In the 2000 presidential election, economic
assessments neither favored nor undermined leftist parties electoral chances. As other scholars
have pointed out, PRIs defeat in 2000 has nothing to do with the economy; on the contrary, the
economic achievements of Zedillos presidency were acknowledged by most Mexicans (Lawson
and McCann 2004). In other words, economic evaluations did not favor leftist parties electoral
chances.
Ideological considerations also play a role in explaining Latin Americans voting

behavior, but this does not mean that voters chose leftist parties at the ballot box because they are
rejecting neoliberalism. Ideological identifications reach significance in Brazil and Uruguay, but
not in Mexico; and policy stances are not relevant voting cues in none of them, with the
exemptions of privatizations and the agrarian reform in Brazil. In order for political cleavages to
become active, politicians need to emphasize them. Torcal and Mainwaring (2003) point out that
130
political cleavages are created by political elites as a way to get votes. The ideological cleavage
only becomes relevant if political leaders and political parties structure political conflict in
ideological terms. Ideology, measured as the individual self placement on the ideological scale, is
a significant predictor of the vote for the left in Brazil and Uruguay: a one unit increase in
ideology (one space to the right on the ideological scale) decreases the probability of voting for a
left-of-center party rather than voting for a center or rightist party. This evidence supports
Hypothesis 2 which states that ideological self-placement is likely to determine the vote for the
left irrespective of social and structural determinants.
On the other side, Mexican politics revolved around a regime cleavage (pro-PRI vs. anti-
PRI) at least until 2000. During that time, the ideological dimension remained inactive, or at
least, as a minor-league dimension (Domnguez and McCann 1995, Domnguez and McCann
1996, Greene 2002, Klesner 2004, Klesner 2005, Magaloni and Poir 2004a, Moreno 1998,
Moreno 1999). Regression results demonstrate that Mexicans ideological self-placement does
not determine their vote. In 2000, individuals who placed themselves on the left side of the
ideological dimension did not significantly differ in their vote from those that placed themselves
on the right.
An alternative way to test the ideological cleavage, and the argument that Latin
Americans shift toward the left is an anti-market response, is to analyze if policy positions are
determinants of voting behavior. To do so, I include a series of variables that measure electorates
opinions towards: state interventionism, redistribution, state regulation of private firms, agrarian
reform, nationalization, and privatizations. The results shown in Table VI-1 indicate that with the
exception of the opinions towards privatizations and agrarian reform in Brazil, none of these
variables are significant determinants to vote for a left-of-center political party. These results
refute Hypothesis 3 which states that Latin Americans who support government involvement
and regulation of the economy, income redistribution and an increase in social spending will be
more likely to vote for leftist parties, while those who are against these policy issues will be more
likely to vote for rightist parties. Despite ideology being a relevant voting predictor, none of the
policy issues traditionally associated with the ideological distinction explains why Latin
Americans choose a leftist party. These results strengthen the argument that Latin Americans
voting behavior is not an ideological rejection of the Washington Consensus.
Latin Americans who had voted left in these pivotal elections do not have a clear
socioeconomic profile. Contrary to the 2006 Latinobarmetro Report which hypothesizes that
poor Latin Americans might be voting left, while wealthy ones will be voting right
(Latinobarmetro 2006), regression results reported in Table 1 dismisses this argument. Only in
Brazil, those with a low family income have a higher probability to vote leftist parties in 1998
and 2002, but this pattern does not exist in Mexico or Uruguay. Neither other socioeconomic
variables like education or the household socioeconomic level reach significance in any of these
countries.
131
Table VI-1. Vote Determinants for leftist Parties: Economic Voting vs. Ideology
Brazil Brazil Mexico Uruguay
Independent Variables: 1998 2002 2000 2004
Economic Voting
Sociotropic Vote 0.344*** 0.416**
(.068) (.182)
Pocket-book Vote 0.686*** 0.369**
(.062) (.191)
Sociotropic Retrospective -0.038 0.054
(.182) (.216)
Pocketbook Retrospective 0.141 -0.037
(.213) (.222)
Sociotropic Prospective 0.403
(.294)
Pocketbook Prospective -0.418
(.284)
Ideology
Ideological selfplacement -0.064** -0.037 -0.798***
(.020) (.048) (.098)
Opinion state interventionism -0.006
(.012)
Opinion redistribution -0.010
(.055)
Opinion state regulations -0.019
(.015)
Opinion agrarian reform -0.431**
(0.141)
Opinion nationalization 0.015
(.015)
Opinion privatizations -0.248*** -0.051
(.048) (.078)
Control variables
Education 0.039 -0.002 0.102 -0.138
(.029) (.017) (.149) (.086)
Family income -.001** -0.001*** 0.094
(.000) (.000) (.064)
Household SES -0.106
(.213)
Household SES (houseware index) -0.109
(.133)
Age 0.001 -0.018*** 0.012 -0.010
(.000) -0.005 (.012) (.008)
Urban Voter -0.167 -0.084 -0.903**
(.157) (.237) (.293)
Partisanship (1)
Left 1.967*** 2.069*** 2.736*** dropped
(.133) (.225) (.445)
Right -0.449*** -0.662*** -2.651*** -2.557***
(.128) (.160) (.431) (.335)
Center 0.065 -0.970
(.124) (1.289)
Constant -2.713*** 2.496*** -0.910*** 3.806**

(0.230) (.594) (.431) (1.505)
Pseudo R squared 0.22 0.18 0.58 0.49
Wald chi2 577*** 181*** 198*** 185520***
Number of observations 2900 1215 710 571
(1) No partisanship is the reference category. *p< .10, ** p< .05, *** p< .01
Note: Entries are binary logit coefficients with robust standard errors
In 2004, STATA dropped party identification with leftist parties from the regression
because they perfectly predict voting for Leftist parties.
132
Partisanship is a strong predictor of voting for leftist parties in these pivotal elections in
Brazil, Mexico and Uruguay; it reaches statistical significance in all of them. Latin Americans
identified with a leftist political party tend to vote for a left-of-center party in presidential
elections, while those identified with a party that belongs to the ideological right also vote within
the same bloc.
To sum up, the evidence from these three country cases show that Latin Americans vote
leftist parties as a way to punish incumbents or traditional parties which were not able to provide
the expected economic well-being. Despite ideology being a relevant voting clue, electorates in
the region are not voting left because they are against the market or strongly in favor of more
state intervention in the economy. They vote for leftist parties because these parties were able to
capitalize the social and economic discontent prevalent in the region. The possibilities of leftist
parties capitalizing on Latin Americans social discontent depend on the number of credible or
untainted opposition. In countries like Brazil and Uruguay, where leftist parties embody the
only credible opposition, it is easy to capture votes from those unhappy with the status quo. But
in countries where more than one credible opposition exists like in Mexico, leftist parties have
to win over the vote of voters who take into account other considerations, mainly the partys
capacity to govern.
Second Argument: Voting Left as a Risk for Latin American Democracy
The recent success of leftist parties in Latin America led some policy analysts to question
the impact that leftist governments could have for democracies in the region. Behind this
question, there is a fear that leftist parties could undermine the stability of democracy in Latin
America. My argument is that the electoral success of leftist parties in Latin America, rather than
being of source of concern, should be taken as an indicator of the democratic maturity of the
region.
Despite the fact that almost all of Latin America emerged from its transition to democracy
some years ago, it still needs to consolidate and improve the quality of its democracy (Roberts
1998). In order for democracies to work properly, they need at least a minimum of popular
support (Easton 1953), and recent research has shown that support for democratic political
institutions and democratic systems depends in part on which side of the winning-losing equation
citizens are (Anderson, Blais, Bowler, Donovan and Listhaug 2005). Citizens that have voted for
a party that lost the election (losers) tend to have lower levels of support for democracy than
winners. As a result, democracies could become unstable if losers are continuously ignored in the
political game, excluded from the political process, and if they are always the same people. To
make democracy strong and stable, according to basic theories of democracy, it is better to have
alternation in power and it is essential to incorporate minorities (Anderson, Blais, Bowler,
Donovan and Listhaug 2005). Following this argument, the recent success of leftist parties in the
region can be taken as good news.
133
In order to test if democracy is in danger due to the success of leftist parties, I test if Latin
Americans who considered themselves leftist differ in their support for democracy from those
that consider themselves rightist or centrist. I hypothesize that:
H4: Latin-Americans who identified themselves with the left are not less democratic than
those who considered themselves centrist or rightist.
Results
To test the previous hypothesis, I use AmericasBarometer 2006-2007 public opinion

survey. This survey covers 23 countries in North America, the Caribbean, Central America and
South America; it is by far the most comprehensive source available to measure democratic
values and behaviors in the Americas. It was carried out during 2006-2007 by LAPOP (Latin
American Public Opinion Project) at Vanderbilt University,9 and it uses a national probability
sample of voting-age adults. For the purpose of this article, I will only use data from Mexico,
Guatemala, El Salvador, Honduras, Nicaragua, Costa Rica, Panama, Colombia, Ecuador, Bolivia,
Peru, Paraguay, Chile, Uruguay, Venezuela, Brazil and Dominican Republic because the focus is
on Latin America; hence I exclude the English-speaking Caribbean and North American cases.
Table VI-2 presents a means comparison analysis of different attitudes and opinions
towards democracy using ANOVA. How strongly citizens defend democracy is measured, in
part, by attitudes toward political tolerance. In the AmericasBarometer we are able to use the
Political Tolerance scale that combines four different indicators10: 1) There are people who speak
negatively of the countrys form of government, not just the incumbent government but the
system of government. How strongly do you approve or disapprove of such peoples right to
vote? 2) How strongly do you approve or disapprove that such people be allowed to conduct
peaceful demonstrations in order to express their views? 3) How strongly do you approve or
disapprove of such people being permitted to run for public office? 4) How strongly do you
approve or disapprove of such people appearing on television to make speeches? The answer
categories go from 1=Strongly disapprove to 10=Strongly approve. Democratic attitudes, or more
political tolerance, correspond to higher values. The index goes from 1 to 100, and it shows a
Cronbachs Alpha of .870 for the pooled sample of countries indicating that all the items have
enough in common to be together.
The Civil Liberties scale is another way to measure how democratic Latin Americans are.
The scale merges five indicators: 1) To what degree do you approve or disapprove of a law
prohibiting public protests? 2) To what degree do you approve or disapprove of a law
prohibiting the meetings of any group that criticizes the nationality political system? 3) To
what degree would you approve or disapprove if the government censored television
9
I want to thank Profesor Mitchell A. Seligson, LAPOP director, for giving me access to this data.
10
When one respondent does not answer one or two of the four questions comprised in the scale, the mean of the
remaining questions is imputed to those questions in order to avoid loosing the observation. However, when more
than two questions are not answered, the case is considered as missing data.
134
programs? 4) To what degree would you approve or disapprove if the government censored
books in public school libraries? 5) To what degree would you approve or disapprove if the
government censored any of the media that criticized it? The answer categories go from
1=Strongly disapprove to 10=Strongly approve. In this case, the scale was reversed to keep the
comparability with the rest of the variables and scales. Therefore, higher values also correspond
with stronger democratic attitudes. The index goes from 1 to 100, and it shows a Cronbachs
Alpha of .788 for the pooled sample of countries11.
TableVI-2. Attitudes towards democracy

Std
Mean N Deviation F Sign.
Political Tolerance scale
Left 56.7 5171 27.988
Center 53.2 7474 25.682 26.126 .000
Right 54.5 6588 28.154
Civil Liberties scale
Left 74.6 4653 21.948
Center 71.4 6523 21.337 38.404 .000
Right 71.1 6125 22.420
Democracy may have problems,
but it is better than any other
form of government (1)
Left 4.99 5119 1868
Center 5.10 7395 1653 71.725 .000
Right 5.36 6498 1.714
Source: AmericasBarometer 2006-2007
(1) Democracy may have problems, but it is better than any other form of government. To what extent do
you agree or disagree with these statements? The answers categories goes from 1= Strongly disagree and
7=Strongly agree.
The results shown in Table VI-2 indicate that leftists in Latin America have, in general,
levels of political tolerance no different from those who consider themselves close to the Right. It
is only those in the political center who have lower tolerance scores. Leftists are stronger
defenders of civil liberties than rightists. However, those who perceived themselves close to the
right, are more in favor of the position that democracy may have problems but it is still the best
form of government than those identified with the left. To sum up, two out of three indicators
support hypothesis 4, which states that Latin Americans who feel themselves close to the left are
not less democratic than those who considered rightist or centrist. In other words, the attitudes of
Latin Americans do not provide strong evidence that the shift towards the left in the region could
imply any danger for democratic prospects.
11
The Civil Liberties scale uses the same procedure to impute missing values than the Political Tolerance scale does.
When one respondent does not answer one or two of the four questions comprised in the scale, the mean of the
remaining questions is imputed to those questions in order to avoid loosing the observation. However, when more
than two questions are not answered, the case is considered as missing data. The N is lower because some of these
questions were not asked in some of the countries.
135
It is possible that the analysis of the pooled sample hides strong country effects that might
contradict the results, but in most countries the differences are not significant. The attitudes
measured by the Civil liberties scale show a similar lack of significance. To make things even
more complicated, Figure VI-1 shows evidence that rightists are more pro-democracy, and have
higher levels of support for the idea that democracy may have problems but it is better than any
other form of government in almost every country with the exception of Chile, Bolivia, Uruguay
and Guatemala. In several of these cases the differences are significant.
Democracy may have problems, but it is
7
better than any other form of
6
government (mean)
5.9 5.9
5.2 5.1 5.5
3 4.6 5.8 4.9 4.9 5.8
6.0
5.7 5.9
5.4 5.5
5.3 5.2 5.1 5.3
5.0 5.0 5.1 5.0
4.7 4.9 4.7 4.9
4.5
4.3 4.8 4.2
2
3.7
V
G
C
M
R
en
lS
an
cu ia
er
os
ol
hi
ol
ra y
ep ela
ua
on
ic as
ru
x
om
le
iv
ar
si
gu
ta
ez
am a
ad
al
.D
te
du
ic
ia
l
ag
va
m
u
o
a
or
b
R
r
om

ic
al
ua
do
.
a
Ideology
Left Right
Source: AmericasBarometer 2006 by LAPOP
Figure VI-1. Democracy is the best form of government by country
Table VI-3 reinforces that last idea: rightists are more or equally democratic than their
fellows on the ideological left12. It is important to mention that the majority of Latin Americans,
in each country, think that electoral democracy is always the best. However, there are some
differences between leftists and rightists. In Costa Rica, Colombia, Nicaragua, Ecuador, Peru,
Venezuela and Panama, rightists are more democratic; while in Chile, Uruguay, Guatemala and
Honduras, leftist are stronger supporters of electoral democracy. Finally, in Dominican Republic,
Brazil, Mexico and El Salvador, the differences are irrelevant.
12
The exact wording of the question is the following: There are people who say that we need a strong leader that
does not have to be elected. Others say that although things may not work, electoral democracy, or the popular vote,
is always the best. What do you think?
136
Table VI-3. Authoritarianism vs. democracy by country

Country Strong leader who does not have to be elected Electoral democracy is the best Total
Brazil
Left 13 87 100
Center 11 89 100
Right 12 88 100
Costa Rica
Left 8 92 100
Center 5 95 100
Right 2 98 100
Chile
Left 4 96 100
Center 8 92 100
Right 16 84 100
Dominican Rep.
Left 13 87 100
Center 10 90 100
Right 12 88 100
Colombia
Left 15 85 100
Center 8 92 100
Right 9 91 100
Mexico
Left 13 87 100
Center 12 88 100
Right 12 88 100
Guatemala
Left 18 82 100
Center 21 79 100
Right 29 71 100
El Salvador
Left 13 87 100
Center 12 88 100
Right 14 86 100
Honduras
Left 14 86 100
Center 12 88 100
Right 16 84 100
Nicaragua
Left 21 79 100
Center 20 80 100
Right 17 83 100
Ecuador
Left 21 79 100
Center 17 83 100
Right 13 87 100
Peru
Left 24 76 100
Center 15 85 100
Right 13 87
Panama
Left 35 65 100
Center 28 72 100
Right 26 74 100
Uruguay
Left 3 97 100
Center 5 95 100
Right 7 95 100
Venezuela
Left 15 85 100
Center 15 85 100
Right 12 88 100
All the countries
Left 15 85 100
Center 14 86 100
Right 13 87 100
Source: AmericasBarometer 2006-07 by LAPOP
137
An alternative way to check if leftists in Latin America are significantly more or less
democratic that rightists is through regression analysis. Table V.4 shows the impact that
ideological self placement has on different indicators of democratic support after controlling by
place of residence (urban versus rural), sex, age, wealth and education. The first thing to notice is
that being leftist or rightist is a significant determinant of all the indicators of democratic support
with the exception of political tolerance. Second, leftists are stronger opponents to government
censorship than rightists. But on the other hand, the evidence is strong in showing that rightists
are more supportive of Churchills idea of democracy (democracy is the best form of
government), and even more important, they are significantly less willing to overthrow
democracy by force. Regression results also indicate that more educated citizens, those with a
higher economic status, and older people show higher levels of democratic support than their
younger, less educated and less well-off fellows. Women, as other studies show, are more
intolerant and more likely to support government censorship than men.
Table V.4 Determinants of Democratic Support

Independent Political Civil Democracy is Opposition to Authoritarianism
variables Tolerance Liberties the best form overthrow an elected vs. Democracy
scale scale of government government (2) (1)
Urban-Rural -.040*** -.052*** -.008 -.026*** .021
Sex -.057*** -.060*** -.010 .036*** .033
Age -.005 .036*** .083*** .089*** .009***
Wealth .100*** -.055*** .142*** .110*** .124***
Education .077*** .136*** .043*** .031*** .028***
Ideological -.008 -.038*** .093*** .071*** .024**
selfplacement
Adjusted R
squared .031 .042 .042 .032 .024
(1) Because Authoritarianism vs. Democracy is a dichotomous variable (0 is Authoritarianism and 1 is Democracy), entries are
binary logit coefficients and the last cell shows the Nagelkerke R Square.
(2) The exact wording of the question is the following: Now we are going to use another card. The new card has a 10-point scale,
which goes from 1 to 10, where 1 means that you strongly disapprove and 10 means that you strongly approve. I am going to read
you a list of some actions that people can take to achieve their political goals and objectives. Please tell me how strongly you
would approve or disapprove of people taking the following actions: of people participating in a group wanting to carry out a
violent overthrow of an elected government. The scale was reversed to keep the comparability with the rest of indicators.
Note: Entries are Beta coefficients from OLS regression models. * p<.10, **p<.05, ***p<.01.
Urban-Rural: Urban 0, Rural 1; Sex: Male 0, Female 1; Age, Wealth and Education: higher values correspond with more age,
more wealth and more years of education. Ideological selfplacement goes from 1 Extreme left to 10 Extreme right. In all the
dependent variables, higher values correspond with democratic attitudes.
Source: AmericasBarometer 2006-2007 by LAPOP.
To sum up, and despite all the country differences found, the results provide evidence to
answer the question if democracy is at risk due to the shift towards the left that the region has
been experiencing. First, it is important to notice that the majority of Latin Americans have
democratic attitudes irrespective of their ideology. This finding is good news for the stability of
democracy in the region. Second, the differences between leftists and rightists on their level of
democratic support depend on the country and the indicator used to measure democratic support.
138
Leftists have higher levels of political tolerance and are stronger supporters of civil liberties than
rightist, but rightists show higher levels of support than leftists to the idea that democracy is the
best form of government in almost all the countries. Even more, rightists are less willing to
overthrow a democratic elected government that their fellows on the left are. To put it simply,
there is evidence that leftists in the region cannot be considered substantially weaker supporters
of democracy than rightists, but it is also not true that they are more democratic than their fellows
on the ideological right.
Conclusion
Taking all these arguments into account, the findings of this article represent a reasonable
positive picture on the prospects for democracy as a result of the rise of the left. First, they show
how Latin Americans have broadened the arena of political contention by electing governments
that incorporate left-oriented parties into the political game. The future of democracy can be in
danger if one important group of political actors are always on the losers side. For many years,
several leftist political parties in the region played as losers in the electoral game. Therefore, the
arrival of left-leaning parties to the government of several Latin American countries, rather than
being a cause of concern, should be considered an indicator of a healthy democracy and a
mechanism to strengthen democratic support among citizens. Second, the results of this study
show that when Latin Americans have institutional and democratic ways to channel their
discontent, they go for them. At least in Brazil, Mexico and Uruguay, voters prefer to vote for
untainted parties rather than looking for non-democratic alternatives to achieve their demands.
In addition, this article supports the idea that Latin Americans are more outcome-oriented
than policy-oriented, and the current shift towards the left in the region is more the result of the
citizens economic discontent than a bandwagon effect against market-oriented economic
policies.
Latin Americans are capable of making their political leaders accountable, remove them
from office when they do not accomplished what was expected of them, and change those in
charge of the government by voting for untainted parties. The success of untainted parties in
Brazil, Mexico and Uruguay implies an increase in institutionalization, in political representation,
and a sign of political maturity (Lpez 2005). In a region demanding a more accountable and
responsive democracy, the examples of Brazil, Mexico and Uruguay show a way this can be
done. The recent shift towards the left in Latin America has helped to intensify and strengthen
democracy in the region by incorporating losers into the political game.
Even though Latin America is better off because of the more pluralist nature of elections,
it is still an open question how the left will behave in office. The analysis performed with the
AmericasBarometer 2006-07 data show that Latin Americans that identified themselves with the
left are not substantially less democratic than rightists, but not more democratic either. Therefore,
in order to figure out where the region is headed as it moves to the left, we need to look at what
leftist governments do: deepening democracy or return to undemocratic political practices. Chile,
Brazil and Uruguay are countries in which leftist governments are acting with respect to
democracy. But the picture is not so clear in countries as Venezuela, Bolivia, and even Ecuador.
139
It is still premature to say if leftists governments in this last group of countries will make
democracy more stable or represent a risk.
140
References
Ames, Barry. The Deadlock of Democracy in Brazil. Ann Arbor: The University of Michigan
Press, 2001.
Anderson, Christopher J., Andr Blais, Shaun Bowler, Todd Donovan, and Ola Listhaug. Losers'
Consent. Elections and Democratic Legitimacy. Oxford: Oxford University Press, 2005.
Benton, Allyson Lucinda. "Dissatisfied Democrats or Retrospective Voters? Economic
Hardship, Political Institutions and Voting Behavior in Latin America." Comparative
Political Studies 38 (4) (2005):417-442.
Cameron, Maxwell. Democracy and Authoritarianism in Peru: Political Coalitions and Social
Change. New York: St. Martin's Press, 1994.
Cantn, Daro and Jorge Ral Jorrat. "Economic Evaluations, Partisanship, and Social Bases of
Presidential Voting in Argentina, 1995 and 1999." International Journal of Public
Opinion Research 41 (4) (2202):413-427.
Converse, P. The Nature of Belief Systems in Mass Publics. In Ideology and Discontent, ed. D.
Apter. Glencoe: The Free Press, 1964.
Domnguez, Jorge I., and James A. McCann. "Shaping Mexico's Electoral Arena: The
Construction of Partisan Cleavages in the 1988 and 1991 National Elections." American
Political Science Review 89 (1) (1995):34-48.
Domnguez, Jorge I. and James A. McCann. Democratizing Mexico: Public Opinion and
Electoral Choices. Vol. 89, American Political Science Review. Baltimore, MD: Johns
Hopkins University Press, 1996.
Dutch, Raymond M. State of the Latin American Political Economy: The James A. Baker III
Institute for Public Policy of Rice University, 2003.
Easton, David. The Political System: Knopf, 1953.
Echegaray, Fabin. Economic Crises and Electoral Responses in Latin America. Maryland:
University Press of America, 2005.
Economist, The. The battle for Latin America's soul. The Economist, 20 May (2006).
Edwards, Sebastian. "Latin America's Underperformance." Foreign Affairs 77 (2) (1997).
Escaith, H. and S. Morley. "El efecto de las reformas estructurales en el crecimiento de Amrica
Latina y el Caribe: una estimacin emprica." El Trimestre Econmico 68 (2001).
Fiorina, Morris P. Retrospective Voting in American National Elections. New Haven: Yale
Fuchs, Dieter and Hans-Dieter Klingermann. The Left-Right Schema. In Continuities in Political
Action, ed. Jennings K. et al. Berlin: Walter de Gruyter (1990).
Gillespie, Charles and L.E. Gonzlez. Uruguay: The Survival of Old and Autonomous
Institutions. In Democracy in Developing Countries, vol. 4: Latin America, ed. Juan J.
Linz Larry Diamond, and Seymour M. Lipset. Boulder, Colorado: Lynne Rienner
Publishers, 1989.
Gonzlez, Luis E. Political Structures and Democracy in Uruguay. Notre Dame: University of
Notre Dame Press, 1991.
Greene, Kenneth. Opposition Party Strategy and Spatial Competition in Dominant Party Regime.
A Theory and the Case of Mexico. Comparative Political Studies 35 (7) (2002):755-783.
141
Huber, Evelyn and Fred Solt. "Successes and Failures of Neoliberalism." Latin American
Research Review 39 (3) (2004).
IDEA. Reform Fatigue. Washington DC: Inter-American Development Bank, 2004.
Inglehart, Ronald and Hans-Dieter Klingerman. Party Identification, Ideological Preference and
the Left-Right Dimensions among the Western Mass Publics. In Party
Identification and Beyond: Representations of Voting and Party Competition, ed. I.
Budge, I Crewey D. Farlie. Chichester: Wiley, 1976.
Kinder, Donald R. Opinion and Actions in the Realm of Politics. In Handbook of Political
Psychology (1998).
Kinder, Donald R. and D. Roderick Kiewiet. "Sociotropic Politics: The American Case." British
Journal of Political Science 11 (2) (1981):129-161.
Kitschelt, Herbert and Staf Hellemans. "The Left-Right Semantics and The New Politics
Cleavage." Comparative Political Studies 23 (2) (1990):210-238.
Klesner, Joseph L. The Structure of the Mexican Electorate: Social, Attitudinal, and Partisan
Bases of Vicente Foxs Victory. In Mexicos Pivotal Democratic Election. Candidates,
Voters, and the Presidential Campaign of 2000, edited by J. I. a. C. L. Domnguez.
Stanford: Stanford University Press, 2004.
Klesner, Joseph L. Electoral Competition and the New Party System in Mexico. Latin American
Politics and Society 47 (2) (2005):103-142.
Kuczynski, Pedro-Pablo and John Williamson, ed. After the Washington Consensus: Restarting
Growth and Reform in Latin America. Washington DC: Institute for International
Economics, 2003.
Latinobarmetro, 2004. Annual Report (2004).
Latinobarmetro, 2006. Annual Report (2006).
Lawson, Chappell and James McCann. "Television News, Mexicos 2000 Elections and Media
Effects in Emerging Democracies." British Journal of Political Science 35 (2004):1-30.
Lewis-Beck, Michael. "Comparative Economic Voting: Britain, France, Germany, Italy."
American Journal of Political Science 30 (2) (1986):315-346.
Lewis-Beck, Michael. . Economics and Elections: The Major Western Democracies. Ann Arbor:
University of Michigan Press, 1988.
Lewis-Beck, Michael, and Paolo Belluci. "Economic Influences on Legislative Elections in
Multiparty Systems: France and Italy." Political Behavior 4 (1982): 93-107.
Lpez, Santiago. "Partidos Desafiantes en Amerrica Latina: Representacin Poltica y
Estrategias de Competencia de las Nuevas Oposiciones." Revista de Ciencia Poltica 25
(2)(2005):37-64.
Lora, Eduardo. Structural Reforms in Latin America: What Has Been Reformed and How to
Measure it. In Working Paper Green Series #348. Washington DC: Inter-American
Development Bank, 1997/2001.
Lora, Eduardo and F. Barrera. Una dcada de reformas estructurales en Amrica Latina: el
crecimiento, la productividad y la inversin, ya no son como antes. Washington DC: Inter
American Development Bank, Research Department, 1997.
Lora, Eduardo, and Ugo Panizza. Structural Reforms in Latin America under Scrutiny. In
Research Paper. Washington DC: Inter American Development Bank, 2002.
Lora, Eduardo, Panizza, Ugo and Myriam Quispe-Agnoli. "Reform Fatigue:Symptoms, Reasons
and Implications." Federal Reserve Bank of Atlanta Economic Quarter (Second Quarter),
(2004).
142
MacKuen, Michael, Robert Erikson and James Stimson. "Sociotropic Politics: The American
Case." British Journal of Political Science 11 (2) (1992):129-161.
Mainwaring, Scott. Rethinking Party Systems in the Third Wave of Democratization. The Case of
Brazil. Stanford, CA: Stanford University Press, 1999.
Mainwaring, Scott and Timothy R. Scully, ed. Building Democratic Institutions: Party Systems
in Latin America. Stanford, CA: Stanford University Press, 1995.
Mora y Araujo, Manuel, and Peter H. Smith. Peronism and Economic Development: The
Elections of 1973. In Juan Pern and the Reshaping of Argentina, ed. F. C. Turner and J.
E. Miguens. Pittsburgh: University of Pittsburgh Press., 1984
Moreno, Alejandro. Party Competition and the Issue of Democracy: Ideological
Space in Mexican Elections. In Governing Mexico: Political Parties and Elections, edited
by M. Serrano. London: University of London, 1998.
Moreno, Alejandro. Ideologa y voto: dimensiones de competencia poltica en Mxico en los
noventa. Poltica y Gobierno 6 (1) (1999):45-81.
Nadeau, Richard and Michael S. Lewis-Beck. "National Economic Voting in U.S. Presidential
Elections." Journal of Politics 63 (1) (2001):158-181.
Panizza, Ugo and Mnica Yaez. "Why are Latin Americans so unhappy about reforms?"
Journal of Applied Economics VIII (1) (2005).
Power, Timothy. "Blairism Brazilian Style? Cardoso and the "Third Way" in Brazil." Political
Science Quarterly 116 (4) (2201/2002).
Przeworski, Adam and John Sprague. Paper Stones. Chicago: University of Chicago Press,
1986.
Remmer, Karen and Francois Glineau. "Subnational electoral choice. Economic and
Referendum Voting in Argentina, 1983-1999." Comparative Political Studies 36 (7)
(2003):801-821.
Roberts, Kenneth M. and Erik Wibbels. "Party Systems and Electoral Volatility In Latin
America: A Test of Economic, Institutional, and Structural Explanations." American
Political Science Review 93 (3) (1999):575-590.
Roberts, Kenneth M. and Moiss Arce. "Neoliberalism and Lower-Class Voting Behavior in
Peru." Comparative Political Studies 31 (2) (1998):217-246.
Rodrik, Dani and Francisco Rodrguez. Trade Policy and Economic Growth: A Skeptic's Guide
to the Cross-National Evidence. In NBER Macroeconomics Annual 2000, ed. B.S. and K.
Rogoff Bernanke. Cambridge: Massachusetts Institute of Technology Press, 2001.
Sachs, Jeffrey; McCord, Gordon and Wing Thye Woo. Understanding African Poverty: Beyond
the Washington Consensus to the Millennium Development Goals Approach. In Africa in
the World Economy, ed. Jan Joost and Age Akkerman Teunissen. The Hague: FONDAD,
2005.
Samuels, David. "Sources of Mass Partisanship in Brazil." Latin American Politics and Society,
2006.
Schlesinger, Arthur M., Jr. The Cycles of American History. Boston: Houghton-Mifflin, 1986.
Seligson, Mitchell, and Miguel Gmez. Ordinary Elections in Extraordinary Times: The Political
Economy of Voting in Costa Rica. In Elections and Democracy in Central (1989)
America, ed. J. Booth and M. Seligson. Chapell Hill: University of North Carolina Press.
Stallings, Barbara, and Wilson Peres. Growth, Employment, and Equity: the Impact of Economic
Reforms in Latin America and the Caribbean. Washington D.C.: Brookings Institution ,
2000.
143
Stiglitz, Joseph. Globalization and Its Discontent. W.W.Norton & Company, 2002.
Torcal, Mariano, and Scott Mainwaring. "The Political Recrafting of Social Bases of Party
Competition: Chile, 1973-95." British Journal of Political Science 33 (2003):55-84.
Weyland, Kurt. "Peasants and Bankers in Venezuela? Presidential Popularity and Economic
Reform Approval, 1989-1993." Political Research Quarterly 51 (2) (1998):341-362.
Weyland, Kurt. "Economic Voting Reconsidered. Crisis and Charisma in the Election of Hugo
Chvez." Comparative Political Studies 36 (7) (2003):822-848.
144
Part II. Challenges to
Democracy from Civil
Society
VII. Social Trust, Economic Inequality, and Democracy

in the Americas *
Abby B. Crdova Guilln**
Abstract
The main task of this research is to identify sources of social trust in Latin America by exploring
why the most politically and economically developed countries in the Americas, namely Canada
and the United States, have higher levels of social trust than their neighbor countries. The major
thesis of this chapter is that the lower levels of social trust in Latin America are in part explained
by the highly unequal distribution of economic resources in many countries in the region because
social trust is unlikely to proliferate in an environment that foster polarization and social conflict
between haves and have nots. The empirical evidence of this chapter suggests that the lack of a
substantial improvement in the distribution of income in Latin America poses a challenge to
democracy given that economic inequality triggers social mistrust. This chapter also finds that
social trust rather than civic participation is the functionally most important component of social
capital in Latin America, since it consistently promotes the type of political attitudes favorable to
democracy, such as support for democracy and political tolerance. This conclusion reinforces the
findings of previous works and suggests that efforts to strengthen civil society in the region are
not likely to advance democracy unless higher civic participation is accompanied by a
considerable increase in the levels of social trust, which in turn seems to require a more
equitable distribution of income.
You cant build trust when some groups feel left out of the society
and believe that others control the resources.1
Is economic inequality eroding Latin Americas social capital? Paradoxically, even though
the emergence of democracy in Latin America has opened important participation channels for its
citizenry, civic and political participation during the democratization period has taken place in an
environment characterized by low social trust, challenging the notion that civic participation and
interpersonal trust, the two main components of social capital, always go hand in hand (Putnam
1993). Taking into account that a high level of social trust has been found in the literature to be
more effective in deepening democracy than the level of participation in voluntary organizations
alone (Inglehart and Welzel 2005), the importance of identifying sources of social trust in order
to strengthen social capital in Latin America speaks for itself. The main hypothesis of this chapter
is that the relatively low levels of social trust in Latin America are in part explained by the highly
* The author is grateful to the comments and suggestions provided by Professor Mitchell A. Seligson on previous
versions of this chapter and to the Department of Political Science at Vanderbilt for its financial support. In addition,
the author would like to thank the members of Latin American Public Opinion Project (LAPOP) team. Any
omissions and errors are entirely the authors responsibility.
**The author is currently a member of the LAPOP team and a PhD student in political science at Vanderbilt
University.
1
Eric M. Uslaner, "Producing and Consuming Trust," Political Science Quarterly 115, no. 4 (2000). p.580.
147
unequal distribution of economic resources in the region, which is the region with the highest
income inequality in the world. In other words, this chapter claims that economic inequality
breeds social distrust, resulting in an overall negative effect on democracy.
Many definitions of social trust can be found in the literature; for example, Newton (2001)
defines social trust as the actors belief that, at worst, others will not knowingly or willingly do
him [her] harm, and at best, that they will act in his [her] interests (202). Generally, social trust
is conceived as the ingredient that makes possible interaction between individuals who might not
know each other well and who do not necessarily share the same interests and values, but who are
willing to undertake collective actions. Therefore, social trust is found to be key for democracy
because it facilitates participation and cooperation among individuals with different backgrounds
(Fukuyama 1995, Inglehart 1999, Putnam 1993, Uslaner 2002), fostering political tolerance.
However, some societies seem more prone to creating social trust than others, and economic
conditions, in particular, appear to play an important role in the formation of social trust.
As the quotation by Uslaner (2000) at the beginning of this chapter suggests, interpersonal
trust is unlikely to flourish in societies where the fruits of economic growth are distributed highly
unevenly since conflicts between the haves and have-nots are likely to be more profound,
undermining solidarity and cooperation and, in turn, weakening the creation of the type of social
capital conducive to better democracies. Although economic growth per se has been found to be
an important determinant of democratic political stability, research has shown that the probability
of democratic breakdown is greater in nations with high levels of economic inequality
(Przeworski 2000). Indeed, much research has pointed to the importance of economic equality in
democratization as well as in conflict prevention (Boix 2003, Karl 2000, Lipset 1961, Muller and
Seligson 1987).
Lipset (1961), for example, stresses the importance of transforming the class structure of
democratic societies from a elongated pyramid to a diamond shape, pointing out that a large
middle class tempers conflict by rewarding moderate and democratic parties and penalizing
extremist groups (51). Indeed, social conflict and personal violence seem to be more common in
societies with a highly unequal income distribution (Fajnzylber, et al. 2002, Fajnzylber, et al.
1998). The argument this chapter offers is that economic inequality hinders the process of
democratic consolidation in Latin America by inhibiting the formation of social trust. Mistrustful
societies are unlikely to create an environment favorable to democracy since low social trust is
associated with greater opportunities for conflict.
In addition, civic participation in an environment characterized by low social trust is likely to

be less meaningful for democracy because membership in civic organizations in this case might
be associated with particularized rather than generalized trust (Uslaner 2002, 2000), which
suggests that trust materializes within but not across civic groups. For this reason, this study
argues that the kind of civic participation that is most important for Latin American democracies
is the one that is accompanied by high levels of interpersonal trust. Thus, this chapter attempts to
determine how Latin American countries could boost their levels of social trust so that the
participation opportunities that have been brought about by the emergence of democracy can be
translated into a widespread democratic culture and therefore stronger democracies. In other
words, how can social capital be made to work in Latin America in a way consistent with the
148
predictions of the bulk of the social capital literature, namely, that interpersonal trust and civic
participation go together and reinforce each other (Putnam 1993). The focus of this chapter has
important policy implications as well as theoretical relevance. Although the strengthening of civil
society has been at the core of democratization programs for the region (Campbell 2003, Seligson
2006), and has been a very important part of USAIDs democracy efforts (Finkel, et al. 2007),
this chapter claims that civic participation by itself is unlikely to foster democracy unless it forms
part of a broader agenda that includes policies that facilitate the conditions for the construction of
generalized social trust, such as economic policies designed to promote equality.
Since both economic equality and socioeconomic development have been identified as
notable sources of social trust, this study also analyzes whether the effect of economic equality
on social trust remains significant once the impact of socioeconomic development is considered.
Indeed, it has been said that feelings of economic security are also likely to produce more trust in
others because economic well-being reduces competition for scarce resources, ameliorating social
conflict (Inglehart and Welzel 2005). Besides structural factors, namely the degree of
development and economic equality, the individual level determinants of interpersonal trust are
also explored. In particular, this work explores the role of personal experience with crime as well
as the standard socio-economic and demographic individual characteristics in the formation of
interpersonal trust. In attempting to identify the determinants of social trust in order to find
mechanisms to strengthen social capital in Latin America, this study departs from previous works
that have considered social trust to be an enduring characteristic, above all a cultural trait, rather
than a social phenomenon that can be shaped through public policy.
Methodologically, a multilevel analysis of the determinants of social trust is carried out,

taking into account individual and country-level characteristics and exploring why it is that, on
average, individuals who live in the more politically and economically developed countries in the
Americas (i.e., Canada and the United States) have much higher levels of social trust than
individuals living in neighbor countries. Indeed, at the country-level, there is overwhelming
evidence, mainly from the World Values Survey data, that there is a strong positive correlation
between social trust and political and economic development (Inglehart and Welzel 2005, Norris
2002). As it will be shown, the empirical evidence presented in this work confirms those results.
The data used in this study come from face-to-face interviews in nineteen Latin American
nations, including Caribbean countries, drawn from the AmericasBarometer in 2006-2007 carried
out by LAPOP, and phone survey data gathered also by AmericasBarometer during the same year
in Canada and the United States. In total, the database includes over 32,000 individual interviews.
The 2006-07 round of surveys facilitated comparisons between two clusters of nations with
dissimilar levels of social trust and political and economic development. Specifically, these data
make it possible to examine why Latin American and Caribbean countries have a lower level of
social trust than Canada and the United States. In order to answer this question, the LAPOP data
are combined with aggregate data about the degree of socioeconomic development and economic
inequality at the country level. The objectives of this chapter are the following:
Briefly review the social capital literature, placing special emphasis on the role of
social trust and its determinants.
149
Measure the level of social capital in the Americas, and determine the strength of
the relationship between social trust and civic participation.
Examine whether the social trust gap between Latin America and Canada and the
United States is determined by differences in personal experience with crime and
socioeconomic and demographic characteristics.
Explore the effect of structural factors, specifically economic inequality and
socioeconomic development, on social trust vis--vis individual-level
characteristics.
Demonstrate the importance of interpersonal trust for democracy in Latin
America, vis--vis participation in civic organizations by examining whether
trusting individuals are more likely to embrace a democratic culture.
A Brief Review of the Literature
The notion that a civic culture is important for the consolidation of democracy has a long
tradition in political science. Social trust has been identified as one of the main characteristics of
a civic culture. For example, Banfield (1958) concluded that Montegrano, a town located in
the south of Italy, was characterized by the almost complete lack of trust among individuals,
which made their cooperative participation in collective enterprises extremely difficult. As a
result, the local government suffered from almost complete political incapacity, which in turn
made improving the economic circumstances of the village impossible. Similarly, Almond and
Verba (1963) found that among the five countries they analyzed, the ones with less stable
democracies, Mexico being one of them, also had lower levels of interpersonal trust.
More recently, the importance of social trust was dramatically reemphasized with the
publication of Making Democracy Work, in which Putnam (1993) made the term social capital
popular. At the heart of the social capital theory rests the idea that democratic governments in
trusting societies are more accountable and responsive to the demands of their citizens because
civic and political participation in those societies is likely to be higher, and therefore
governments are, at the same time, more likely to represent the interests of the bulk of the
population. Putnam suggests that the main components of social capital are interpersonal trust
and civic participation. In fact, for Putnam there is a reciprocal relationship between interpersonal
trust and civic participation, suggesting that where civic participation is high, social trust should
also be high, and vice versa. However, as it will be shown in this chapter, although civic
participation is relatively high in many Latin American countries, the level of social trust is low,
empirical evidence that contradicts Putnams theory.
The common characteristic among the aforementioned works is that they understand political
outcomes to be products of a civic culture that is rooted in the historical legacy of nations and
therefore in the cultural values transmitted from generation to generation. From this perspective,
social trust is predominantly a cultural phenomenon, making changes in the levels of social trust
150
difficult since by their very nature, cultural traits are enduring.2 This also implies that some
cultures are more favorable to democracy than others.
Following Max Webers ideas, Western Protestant values are seen more compatible with
democracy than other value systems. Indeed, many of the politically and economically successful
countries in the world are at the same time those whose cultures have been rooted in Protestant
traditions Even though the so-called Confucian ethic civilizations have made impressive gains
in recent years. For example, Harrison (1985) evaluates the causes of underdevelopment in Latin
America, and concludes that what impedes prosperity in Latin American countries are Catholic
values and traditions imbedded in the Hispanic culture. These views are shared by Inglehart
(Inglehart 1990, 1988, 1999, Inglehart and Welzel 2005), who understands civic attitudes,
including social trust, as cultural products, although he acknowledges that structural factors,
socioeconomic development in particular, can shape the levels of social trust specially in already
politically and economically advanced societies.
These culturalist approaches have been challenged by the empirical evidence, since, for
example, social trust and civic participation appear to have experienced important declines in the
United States in recent decades (Putnam 2000, 1995), suggesting that some other factor besides
deep-seated cultural traits might be affecting the levels of social capital in the U.S., given that
dramatic cultural changes are unlikely to take place within a relatively short period of time.
Putnam argues that the decline in civic activism in the United States is explained, in part, by the
increased number of hours people spend watching television, which deter individuals from
interacting with others. In addition, he claims that television viewing has also made people less
trusting because television programs tend to overstate negative aspects and portray the world as a
mean place (Putnam 1996).
The Putnam thesis has not gone unchallenged. While many agree that there has been a
decline in social trust in the U.S., the explanation is not to be found in television viewing.
Uslaner (1998, 2005), for example, rejects Putnams explanations for the decline in social capital
in the United States. Instead, he argues that the recent rise in economic inequality in the United
States better explains the waning of interpersonal trust (Uslaner and Brown 2005), and adds that
the decline in civic participation is the effect of the decline in social trust.
Indeed, other authors have also suggested that the elements that make up a civic culture,
among them social trust, are not entirely cultural, but the byproduct of structural factors (Armony
2004, Jackman and Miller 1998, Muller and Seligson 1994). Therefore, civic attitudes are more
malleable than what some proponents of the culturalist view suggest. Structural factors are
important, and consequently social trust can be shaped by the products of public policies. This is
the theoretical framework that has inspired this study. Regardless of the role that culture might
play, this study argues that the quality of social capital in Latin America can be improved if and
only if the appropriate set of public policies is implemented. Economic equality has been found
to be an important structural factor that affects the level of social trust in democratic nations
2
Cultures can change over time, as Inglehart has demonstrated in his studies of Europe. Ronald Inglehart, Culture
Shift in Advanced Industrial Society (Princeton, N.J.: Princeton University Press, 1990).
151
(Armony 2004, Uslaner and Brown 2005), and its role in shaping Latin Americas social capital
vis--vis other structural and individual factors is investigated here.
Social Capital in Latin America in Comparative Perspective
Following Norris (2002)s approach, this study measures social capital using two
indicators: social trust and civic activism.3 The following item in the AmericasBarometer survey
measures the level of interpersonal trust:
IT1. Now, speaking of the people from around here, would you say that people in this community are
generally very trustworthy, somewhat trustworthy, not very trustworthy or untrustworthy?
Social trust, in this study, then, refers to how confident individuals feel about people outside
their immediate family circle. As mentioned before, the type of trust that is likely to enhance
democracy is the one that makes individuals sympathetic to people who may have different
beliefs, ideologies, or backgrounds than ones own, and consequently this type of trust ought to
contribute to the establishment of more harmonious and, ultimately democratic societies (Putnam
2000, Uslaner 2002). Figure VII.1 shows the average level of social trust in each country, based
on AmericasBarometers measure of interpersonal trust.4 As expected, the average level of social
trust is much lower in Latin American and Caribbean countries than in Canada and the United
States, confirming the trend found in the World Values Survey data. The two richest and more
democratic countries in the sample show a much higher level of social trust. Out of the twenty
one countries examined, Canada has the highest level of social trust, with an average of 82.8
points, and Haiti has the lowest level, with an average of only 42.1 points. The United States has
the second highest level of social trust, a level that is not statistically significantly different to that
of Canada.5
3
Factor analysis of the AmericasBarometer 2006 dataset revealed that the measures of social trust and civic activism
here used belong to different dimensions, which suggest that each of them is a distinct component of social capital,
confirming Norris (2002) findings. Norris states that social trust is the cultural dimension of social capital, and
civic activism the structural dimension. In contrast to Norris, however, a single index of social capital was not
constructed because a Cronbach Alpha test revealed that there is a low reliability of an index constructed using both
of these components, which suggest that, at least in the Americas, social trust and civic participation do not always
go together.
4
The original interpersonal trust variable in the survey was recoded into a 0-100 scale in order to retain
comparability with the other chapters in this volume, and therefore facilitating cross-chapter comparisons.
5
Although the AmericasBarometer 2006 included data for Brazil, this country was excluded from the analysis in this
chapter because the wording of interpersonal trust variable is different.
152
Canada 82.8
United States 79.6
Honduras 67.2
Costa Rica 67.0
Paraguay 65.6
Colombia 62.7
Uruguay 62.6
El Salvador 62.2
Dominican rep. 60.4
Nicaragua 60.2
Guyana 59.9
Guatemala 59.1
Chile 58.9
Jamaica 58.9
Mexico 58.6
Venezuela 53.9
Ecuador 53.9
Panama 49.4
Bolivia 47.0
Peru 43.0
Haiti 42.1
0 20 40 60 80 100
Average Social Trust
Source: AmericasBarometer 2006 Error Bars: 95% CI
Figure VII-1. Average Social Trust by Country
In order to measure the level of civic activism, an index of civic participation was
constructed, which indicates whether a respondent stated having participated in meetings of at
least one of the following secular voluntary organizations: a parents association at school (the
equivalent of the U.S. PTA), a committee or council for community improvement, an association
of professionals, an association of merchants or farmers, and meetings of a political party or
movement.6 Figure VII.2 shows the percentage of the population in each country that participates
in at least one of these four types of civic organizations. The three countries with the highest level
of civic participation, based on the index here developed, are Bolivia, Peru, and Canada, with
6
Although the survey included an item about participation in meetings of religious organizations, it was not included
in the index because factor analysis shows that participation in this type of organizations belongs to a different
dimension or component of civic participation. It is likely that respondents to this item include their participation in
religious services as part of their count of participation, and the analysis shows this kind of activity to be distinct
from other forms of civil society activism. In addition, although the survey asked how frequently respondents
participated in the meetings of all these organizations, a dichotomous index, instead of a scale, was preferred,
because the reliability of the four items included in the index of civic participation was low. Low reliability means
that the organizations do not naturally group together, which is understanable; members of farmer associations are
not likely to be members of professional associations.
153
78.2, 73.8, and 72.4 percent of the population, respectively, participating in the civil society. In
sharp contrast, the countries with the lowest percentage of the population participating in
meetings of civic organizations are Panama (45.4), Costa Rica (46.5), and Uruguay (47.2). On the
face of it, at least, these results do not conform to expectations.
While it was expected that Canada would show a high level of civic participation, it was also
expected that the U.S. would also be at the top, but it was not. Nonetheless, in absolute terms,
participation in the civil society in the U.S. is respectably high (67.5%). The real surprises are at
the top and bottom for Latin America. Panama, Uruguay, and Costa Rica are well established
democracies, with the latter widely considered to be the oldest and most consolidated democracy
in the region. Yet, these three nations score at the very bottom. Similarly, Chile, in many ways a
leading example of a successfully re-democratized country, is also low. On the other hand,
Bolivia scores at the very top, a country with a fragile democracy and a turbulent democratic
history in which elected regimes have frequently been forced from office.
Bolivia 76.2
Peru 73.8
Canada 72.4
Haiti 71.2
Dominican rep. 70.5
Guyana 68.4
Jamaica 68.3
United States 67.5
Ecuador 65.1
Honduras 63.8
Mexico 60.1
Venezuela 59.5
Colombia 59.4
El Salvador 58.2
Nicaragua 54.0
Chile 53.2
Guatemala 51.4
Paraguay 48.0
Uruguay 47.2
Costa Rica 46.5
Panama 45.4
0 20 40 60 80
Percentage who participate at least in one civic organization
Fuente: AmericasBarometer 2006 Error Bars: 95% CI
Figure VII-2. Civic Participation in the Americas
154
Figure VII.3 clusters the countries analyzed into regions of high and low social capital,
namely Latin American and Caribbean countries and Canada and the United States.7 At the
regional level, social capital, measured by social trust and civic participation, is higher in Canada
and the United States than in Latin America, although social capital differentials between these
two regions are more substantial for social trust. While the civic participation differential
between the means of the two regions is only 9.8 points, the difference in social trust is 23.7
points. As observed in the previous figures, in some Latin American countries civic activism is
higher than in Canada and the United States, but social trust is significantly lower in Latin
America. Overall, the less politically and economically developed regions in the Americas show
lower levels of social trust.
100
United States and
Canada
Latin America and
the Caribbean
80
60
Average
40 81.2
70.0
60.2 57.5
20
0
Participate in at least one Social Trust
civic organization
Figure VII-3. Social Capital by Region
A Weak Link: Social Trust and Civic Participation
In this section, we assess the proposition that social trust and civic activism form a virtuous
circle given that these two elements of social capital have been thought to reinforce each other,
and therefore it is assumed that high social trust leads to participation in civil society, and vice
7
Herein, the term Latin America in this chapter is used to refer to all countries in the survey, except Canada and
the United States.
155
versa (Putnam 1993). As noted in the introduction of this chapter there are reasons to question
this putative link, especially as Armony (Armony 2004) has shown that participation very
frequently does not lead to trust (or democratic outcomes). He notes the extremely high levels of
participation in some areas in Weimer,Germany, precisely the areas in which the Nazis first
gained control. The AmericasBarometer data reinforce this doubt.
Figure VII.4 shows the relationship between interpersonal trust and civic participation at the
country level. It can be observed that the relationship between interpersonal trust and civic
participation is not well defined. For example, while Costa Rica has a relatively high level of
interpersonal trust in comparison to other Latin American countries, it also has one of the lowest
levels of civic participation. On the other hand, the Latin American countries with the highest
civic participation (Bolivia, Peru, and Haiti) show the lowest levels of interpersonal trust. Only
Canada and the United States show a relatively high average level of social trust and civic
participation. This suggests that at the country level there is no clear-cut association between
social trust and civic participation, as Putnam (1993)s social capital theory suggests that there is.
This finding is not entirely novel, however, as other survey-based studies have also found that
social trust and civic participation do not go always together (Newton 2001, Norris 2002,
Seligson 2002).
80
Bolivia
Percentage who participate in at least one civic
Peru
Canada
Dominican Rep.
70 Haiti
Jamaica United States
Ecuador Guyana
Honduras
organization
Mexico
60 Venezuela Colombia
El Salvador
Nicaragua
Chile
50 Guatemala Paraguay
Panama
Uruguay
Costa Rica
40
40 50 60 70 80 90
Figure VII-4. A Weak Link: Social Trust and Civic Participation
156
In light of these finding, it must be asked, at the individual level, if a higher level of
interpersonal trust leads to, or is at least associated with, higher levels of civic activism.
Figure VII.5 again shows that this seems to be the case only in Canada and the United States
where individuals who have a higher level of interpersonal trust also show a higher average
level of civic participation. In Figure VII.5, the line that represents the Northern countries has
a steep slope, indicating that social trust is very closely linked to civic participation. In
contrast, in the Latin American region, higher levels of social trust do not seem to be
associated with higher levels of participation in the civil society, as reflected by the almost
flat line in the figure. While in Latin America about 60 percent of the individuals who have
the highest level of social trust participate in at least one type of civic organization, in Canada
and the United States over 70 percent do. This supports Norris (2002) observation that the
contextual (i.e., national) level of social trust matters for civic participation. Even if the
individual level of interpersonal trust is high, this only results in civic activism if other
members of society also show high levels of trust.
75
United
States
and
Canada
70 Latin
America
and the
Caribbean
65
organization
60
55
50
Untrustworthy Not very Somewhat Very

trustworthy trustworthy trustworthy
Social Trust
Figure VII-5. The Link between Social Trust and Civic Participation: and individual level analysis
However, Figure VII.5 also indicates that in Latin America even individuals who have no
confidence in others at all do indeed participate in meetings of secular civic organizations. A
larger percentage of respondents in Latin America with the lowest level of social trust are
members of civic organizations than in Canada and the United States. This suggests that in Latin
157
America civic organizations may not bring together people of different beliefs, ideologies, or
backgrounds, since social trust facilitates interaction between dissimilar individuals. Indeed, it
could be argued that the benefits of civic participation for democracy are more likely to be
maximized in heterogeneous rather than homogenous organizations since plural civic
organizations have been found to foment political tolerance (Theiss-Morse and Hibbing 2005). In
general, the results suggest that other factors seem to be more important in Latin America for
bringing about civic participation than social trust.
In fact, at the individual level, in Latin America the level of education proved to be a more
important predictor of civic participation than social trust. Indeed, the effect of social trust on
civic participation in Latin America, although positive and statistically significant, is negligible.8
Figure VII.6 shows that there is a relatively strong relationship between the individual level of
education and civic participation. This confirms the uneven character of social capital. Those who
have more human capital are more likely to join voluntary organizations, since they have the
means (i.e., resources, in Verbas terminology) to participate more effectively (Putnam 2000,
Verba, et al. 1995). However, Figure VII.6 also shows that a substantial percentage of persons
without formal education in Latin America, over 50 percent, do participate in meetings of secular
civic organizations.
65
62
organization
60
58
55
52
None Primary Secondary Tertiary
Source: AmericasBarometer 2006 sig.<0.001
Figure VII-6. Level of Education and Civic Participation in Latin America and the
Caribbean
8
The results of the regression analysis can be found in appendix VII.1. The design effect of the sample was taken
into account for the estimation of the standard errors of the coefficients in all the regressions in this chapter.
158
Explaining the Social Trust gap in the Americas: Do Individual Characteristics

Matter?
This section explores whether the lower level of social trust in Latin America relative to
Canada and the United States is explained by individual level factors, such as the varying socio-
economic and demographic characteristics and personal experiences. Levels of social trust in
Latin America and Canada and the United States are compared taking into account the individual
characteristics of respondents. If individuals with similar characteristics, but living in very
dissimilar social contexts, show different levels of social trust, it would suggest that structural
factors might be more important in explaining social trust differentials in the Americas than
individual characteristics.
Socio-economic Characteristics
Do education differentials explain the lower level of social trust in Latin America and the
Caribbean? It is well documented in the literature that more educated individuals show higher
levels of social trust (Newton 2001, Putnam 1995). As a result, one hypothesis that emerges is
that the lower level of social trust in Latin America is explained by lower levels of schooling in
this region. If this were the case, it might be observed that individuals with very low levels of
education in Latin America also have lower levels of social trust while well educated individuals
would show levels of social trust similar to that of Canada and the United States. Figure VII.7
clearly rejects this hypothesis.
159
100
Canada and the
United States
Latin America and
Average Social Trust 80 the Caribbean
60
40 83.1 79.5 82
58.9 59.8 58.1

55.5
20
0
Figure VII-7. Social Trust by Educational Level
Schooling does not seem to explain the lower level of social trust in Latin America. There
are no significant differences in the average level of trust across educational levels in the two
regions. And, the gap in social trust between Latin America and Canada and the United States
remains large at each level of education. In fact, the social trust differential is still about 23 points
at each educational level, similar to the overall regional gap observed in Figure VII.3.
Demographic Characteristics
Are demographic characteristics responsible for the lower level of social trust in Latin
America? Specifically, the first issue is whether men and women have similar levels of social
trust in Latin America and Canada and the United States or if there is a gender bias in social
capital that explains social trust differentials across regions. Second, it must be determined
whether there are differentials in the age composition of the population in these two regions that
would explain the social trust gap. Indeed, previous works have found that that in general social
trust seems to vary by sex and age. There is evidence that men and middle age persons tend to be
more trusting (Newton 2001), and therefore gender differences, along with differentials in the age
composition of the population are also plausible factors explaining why the two sets of countries
here analyzed show different levels of social trust. The following two figures help to answer
these questions.
160
Figure VII.8 shows that, on average, there are no substantial differences in the level of social
trust between men and women either in Latin America or in Canada and the United States.
Therefore, the gap in social trust between regions is still evident even when the sample is split
across genders.
100
Canada and the
United States
Latin America
80 and the
Caribbean
60
40 81.1 81.2
58.8 56.2
20
0
Male Female
Figure VII-8. Social Trust by Gender
Figure VII.9 shows that the age composition of the population in the two regions is not
the factor explaining the social trust gap either. In both regions, the trend is similar: older
people are, in general, more trusting. The differentials in social trust across regions remain
constant at about 20 points in each age cohort.
161
100
Canada and the
United States
Latin America
and the
80 Caribbean
60
86.5
40 81 80.6 80.7
75.8 75.8
63.5
57.8 59.2 60.1
54.8 56.6
20
0
16-25 26-35 36-45 46-55 56-65 66+
Age
Figure VII-9. Social Trust by Age
Personal Experience with Crime
Finally, this section evaluates whether the higher rate of crime victimization in Latin America
relative to Canada and the United States explains the social trust gap between these two groups of
countries. The 2006-07 AmericasBaromenter data on individual crime victimization show that, as
expected, on average, crime rates are higher in Latin American countries than in Canada and the
United States. While in Latin America, according to the AmericasBarometer survey, the
percentage of the population victimized by crime is 17.5, in Canada and the United States that
figure is 11.7 percentage points. Since individuals who are victims of crime might be less likely
to trust others in their communities, the higher crime victimization rate in Latin America is a
plausible candidate to explain the social trust gap in these two regions. In fact, there is evidence
that social trust and crime victimization are negatively correlated (Lederman, et al. 2002). Figure
VII.10 shows, however, that differentials in crime victimization rates do not seem to explain the
social trust gap.
162
100
Canada and
the United
States
80 Latin America
and the
Caribbean

60
40 82.1
73.8
58.6
52.7
20
0
No Yes
Victim of Crime
Figure VII-10. Crime Victimization and Social Trust
Although, as expected, on average those who reported having been victims of crime show a
lower level of interpersonal trust, this does not explain the differential in social trust between the
two regions analyzed here. If crime victimization were an important factor explaining social trust
differentials, similar levels of social trust could be found in both regions among those who did
not have a personal experience with crime; however, clearly this does not seem to be the case.
The average level of social trust is lower in Latin America, regardless of whether a person has
been a victim of crime or not, even though in both regions trust is lower among those who have
been victimized by crime. Since we know that criminals do not seek out their victims based upon
the latters level of trust, we know that low trust is a product of being a victim rather than the
other way around.
In conclusion, this section did not find empirical support for the hypothesis that socio-
economic and demographic individual characteristics or personal experience with crime explain
why in Latin American countries individuals are less trusting of people outside their family group
than in Canada and the United States. Without further analysis, the evidence presented so far
seems to lend support to the thesis that deep-seated cultural traits might be important
determinants of the type of a civic culture that one finds in different regions of the world.
Although this study does not test the culturalist thesis directly, the following section provides
evidence that structural economic factors explain, at least in part, why in the Americas some
societies are more trusting than others.
163
Economic Inequality and Social Trust: A Multi-level Analysis
Until now, the analysis performed in this study has shown that the Latin American region
unambiguously has a lower average level of social trust than the much more economically
developed and democratic Canada and the United States; moreover, we have observed that the
social trust gap in these two regions does not seem to be explained by gender biases, differentials
in schooling, the age composition of the population, or differentials in the level of crime
victimization across regions. The next step in the analysis is to determine whether the context
matters, and if so, what structural factors in these countries seem to explain such social capital
differences. In this section, we test the hypothesis that countries with higher income inequality
show lower levels of social trust, or in other words, that individuals living in countries
characterized by a particularly unequal distribution of income are more likely to distrust people in
their communities.
The study of the consequences of economic inequality for social trust, and therefore
democracy, is especially important for Latin America, which has the most unequal income
distribution compared to other world regions, and the lack of equality is particularly notorious
when it is compared to that of the developed world. According to recent data from the World
Bank, whereas the richest tenth of the population in the region earn 48 percent of total income,
the poorest tenth earn only 1.6 percent. By contrast, in developed countries the top tenth receive
29 percent of total income, compared to 2.5 percent for the bottom tenth.9 In fact, although
income inequality has been rising in the United States in recent years, it still has a lower level of
inequality than any of the Latin American and Caribbean countries examined in this chapter. This
study argues that the great economic inequality in Latin America is hindering social capital by
lowering the overall level of social trust.
In order to test this hypothesis, we estimated a Hierarchical Linear Model (HLM) which takes
into account the fact that individuals are nested within countries. The properties of these
statistical models make it possible to assess how individual characteristics affect the dependent
variable, in this case the individual level of interpersonal trust and at the same time whether
structural characteristics shape individuals attitudes.10 In other words, this statistical tool allows
us to explore why, once individual characteristics are controlled for, an individual living in a
Latin American country shows a lower level of social trust than an individual with similar
characteristics living in the United States or Canada; that is how the context affects the average
level of social trust in a country. As mentioned before, the main claim of this chapter is that
income inequality is an important structural variable that explains, at least in part, the lower level
of social trust in Latin American countries.
At the individual level, four variables were included in the model: years of schooling, gender,
age, and crime victimization. These are the same variables analyzed in the previous section using
data from the AmericasBaromenter surveys. At the country level, two competing structural
9
David M. De Ferranti, Inequality in Latin America : Breaking with History?, World Bank Latin American and
Caribbean Studies. Viewpoints (Washington, DC: The World Bank, 2004). p.1-1.
10
For an overview of HLM, see Douglas A. Luke, Multilevel Modeling (Thousand Oaks, Calif.: Sage Publications,
2004).
164
variables were tested: income inequality and socioeconomic development. The economic
inequality data comes from the World Bank Development Indicators, and correspond to the most
recent estimate of the Gini coefficient available for each country in the sample.11 The
socioeconomic development variable corresponds to the UNDP Human Development Index for
2004, the most recent available.12
Table VII.1 summarizes the results of the multilevel model.13All the variables included in the
model are statistically significant and show the expected effect on interpersonal trust. On average,
males, more educated, and older individuals show a higher level of interpersonal trust, while
victims of crime are more likely to distrust people in their communities. In addition, the results
show that social trust is higher in more developed countries and in those where income inequality
is lower.
Table VII-1. Structural and individual predictors of social trust: Hierarchical Linear Model
Variables Coef. Standard Err.

Structural (country-level)
Income inequality (Gini) -0.101** 0.052
Socio-economic development (HDI) 0.116** 0.051
Individual-level
Years of Schooling 0.040*** 0.006
Male 0.092*** 0.011
Age 0.080*** 0.006
Victim of crime -0.195*** 0.014
Constant 0.003 0.052
Note: *Standardized coefficients for continuous variables
**p-value<0.05
***p-value<0.001
Number of observations: 31.859
Number of groups (countries): 21
R-squared (total): 5.5%
R-squared (structural variables): 5.3%
R-squared (individual variables): 0.2%
Structural factors, economic inequality and socioeconomic development, on the other hand,
are much more important predictors of social trust than individual level factors. This conclusion
was drawn because although the results revealed that 5.5 percent of the total variance is explained
by structural and individual factors, about 5.3 percent of the total variance is explained by
11
The Gini coefficient can have a value between zero and one; it measures the extent of income inequality in a
country. A larger Gini indicates a greater degree of income inequality.
12
The Human Development Indicator (HDI) is a composite measure of the level of socioeconomic development in a
country. It encompasses three measures of socioeconomic well-being: economic resources (GDP per capita at
purchasing power parity), an index of education, and a health indicator, as measured by life expectancy at birth. The
HDI ranges between zero and one, with a higher value denoting a higher level of development.
13
Appendix VII.2 shows the decomposition of the variance of our social trust variable. The results show that about
10% of the total variation in social trust is due to cross-country differences. This relatively small variation across
countries is explained by the fact that although there is a substantial difference in the average social trust between
Latin American and Northern countries (Canada and the United States), this difference is being offset by the smaller
differences in social trust across Latin American countries.
165
structural factors alone. 14 Moreover, the statistical analysis shows that socioeconomic
development and economic inequality have almost the same effect on social trust (although in
contrary directions) as shown by the absolute value of the coefficient for each of these variables.
This finding reinforces the results shown in the previous section; individual characteristics are
less important in explaining interpersonal trust differentials. This means that the most important
sources of social trust are neither personal characteristics nor experience with crime, but rather
the economic context of the place of residence. Figures VII.11 and VII.12 show the relationship
between social trust and income inequality, and socioeconomic development at the country-level,
accordingly.
90
Canada
The United States
80
Honduras
70 Costa Rica
Paraguay
Uruguay El Salvador
Jamaica Nicaragua Dom. Rep Guatemala Colombia

60
Guyana Mexico Chile
Ecuador
Venezuela
Panama
50 Bolivia
Peru
Haiti
R Sq Linear = 0.375
40
30 35 40 45 50 55 60
Economic Inequality (Gini coefficient)
Source: AmericasBarometer 2006; WB Development Indicators
Figure VII-11. Economic Inequality and Social Trust
As can be observed in Figures VII.11 and VII.12, inequality and development are two
forces working in opposite directions that affect social trust. This suggests that although
improving the socioeconomic development of Latin American countries is important,
improvements in the standard of living of the population alone are not likely to result in a
significant increased in social trust, and therefore in stronger democracies, unless an
improvement in the distribution of income also takes place.
14
Although the explanatory power of the model seems to be small, since it only explains 5.5 percent of the total
variance, it is important to keep in mind that regression analysis based on survey data usually yields low R-squared
values.
166
90
Canada
80
The United States
Paraguay
70 Honduras Costa Rica
El Salvador Colombia
Nicaragua
Dom. Rep Uruguay
60 Guatemala
Guyana Mexico Chile
Jamaica
Venezuela
Ecuador
Panama
50 Bolivia
Peru
Haiti
R Sq Linear =
0.421
40
0.4 0.5 0.6 0.7 0.8 0.9 1.0

Human Development Index
Source: AmericasBarometer 2006; United Nations (UNDP) 2004
Figure VII-12. Socioeconomic Development and Social Trust
Furthermore, the findings of this section suggest that efforts to strengthen the rule of law
and to mitigate social violence in Latin America are more likely to have a positive impact on
democracy if at the same time the sources of conflict are mitigated. The great economic
inequality faced by most Latin American countries constitutes a source of social conflict. Indeed,
as mentioned before, prior research has found that social violence is more common in countries
with higher economic inequality (Fajnzylber, et al. 2002, Fajnzylber, et al. 1998), and the results
of this study suggest that a possible intervening factor explaining that outcome is the lower levels
of social trust found in highly unequal societies. Therefore, in order to reduce social violence, an
increase in social trust seems necessary, which in part can be achieved through a better
distribution of wealth.
The Importance of Social Trust for Latin American Democracies
In the previous section, we identified sources of social trust in the Americas. In this
section, we demonstrate that increasing the levels of social trust is imperative for fomenting a
widespread democratic culture in Latin America. In general, we expect that individuals with a
higher level of interpersonal trust will show a higher support for democracy, and we also expect
them to be more willing to respect the civil liberties and political rights of minorities or of those
whom they are more likely to disagree with. A democratic culture has been argued to be a
167
prerequisite for the consolidation of democracy (Diamond 1999), and support for democracy and
political tolerance are indispensable values of highly democratic individuals. Interpersonal trust
encourages those values. Trusting individuals are expected to oppose authoritarian rule since they
are more likely to feel confident in cooperating with others and peacefully accept political
outcomes even when such outcomes do not directly favor them. Indeed, the empirical evidence
suggests that trusting individuals have a higher probability to tolerate and support policies that
benefit minorities and the poor (Uslaner 2004). Therefore, trusting individuals are also likely to
have a higher political tolerance because they are more willing to interact with people unlike
themselves, and consequently, more likely to respect the rights and worldviews of others.
Figure VII.13 demonstrates that, in general, the higher the individual level of
interpersonal trust, the higher the average support for democracy in Latin America. The
significant positive relationship between interpersonal trust and support for democracy holds
even after taking into account the individuals evaluation of the economic situation of the
country, their satisfaction with the way democracy works, their approval of current government,
participation in civic organizations, and individual demographic and socioeconomic
characteristics.15
75
Average Support for Democracy
72
70
68
65

Level of Social Trust
Source: AmericasBarometer 2006 sig.<0.001
Figure VII-13. Support for Democracy and Social Trust in Latin America and the Caribbean
15
The regression results for support to democracy are show in Appendix VII.3. The support for democracy
variable corresponds to the ING4 variable in the AmericasBarometer survey which asks individuals to state the
extent to which they agree with the following proposition in a scale from 1 to 7: Democracy may have problems,
but it is better than any other form of government. To what extent do you agree or disagree with this statement?
168
In order to test the proposition that higher interpersonal trust is associated with higher
political tolerance at the individual level, the political tolerance index developed by Mitchell
Seligson of LAPOP was included as dependent variable in the multivariate statistical analysis.16
Figure VII.14 shows that there is a positive relationship between interpersonal trust and political
tolerance. This association remains significant even after controlling for the effect of civic
participation, approval of current government, and other individual characteristics.
Besides showing that higher interpersonal trust translates into higher support for democracy
and political tolerance, the statistical analysis revealed that interpersonal trust is consistently the
most important component of social capital in Latin America, a finding that supports the recent
world-wide study of Inglehart and Welzel. When the effect of civic participation on support for
democracy and political tolerance are examined, the only mode of civic participation that has a
positive effect on support for democracy is participation in meetings of religious organizations;
however, participation in such meetings is at the same time associated with lower political
tolerance. On the other hand, although participation in meetings of a political party or movement
is the only mode of civic participation likely to significantly increase political tolerance,
paradoxically, activism in political organizations has a negative effect on support for democracy.
60
58
Average Political Tolerance
56
54
52

Level of Social Trust
Source: AmericasBarometer 2006 Sig.<0.001
Figure VII-14. Political Tolerance and Social Trust in Latin America and the
Caribbean
16
The regression results for political tolerance are shown in Appendix VI.4. The political tolerance index is based on
items D1, D2, D3, and D4 in the survey.
169
Conclusion
This chapter has demonstrated that social trust is the most important component of social
capital in the Americas; when civic participation does not take place in an environment
characterized by generalized social trust, as it seems to be the case in Latin America, civic
activism is unlikely to contribute to the configuration of a democratic culture. In addition, this
study, based on the AmericasBarometer data, confirmed the trend found in previous studies; the
most politically and economically developed countries in the sample, namely Canada and the
United States, show the highest levels of social trust, which reinforces the importance of social
trust for development. For this reason, the main task of this chapter was to identify sources of
social trust, and therefore explain why Latin American countries have a lower level of social
trust. This chapter showed that, in general, individual characteristics and personal experience
with crime have a significant impact on the individual level of interpersonal trust, and that these
individual factors are not responsible for the social trust differentials between Latin America and
Canada and United States. The statistical analysis showed that the most important predictors of
social trust are structural rather than individual factors, namely economic inequality and
socioeconomic development.
Indeed, the results show that even if the effect of economic development is taken into
account, economic equality remains a significant source of trust. Specifically, the findings
suggest that economic development is likely to enhance democracy by increasing the levels of
social trust as long as there is a substantial improvement in the distribution of economic
resources. A conclusion that gives support to the thesis that although economic development
fosters democracy, individuals are likely to embrace the kind of values necessary to support
democracy only if socioeconomic development is also accompanied by economic equality
(Lipset 1961, 1959).
Taken together, these results suggest that although Latin American countries, in general,
have progressed in some areas, reflected for example in the increased years of schooling in the
region (IADB 2006), the impact of these advances on democracy is perhaps offset by the lack of
a substantial improvement in the distribution of income in the region. Therefore, fiscal reform
that seeks redistribution is imperative for democracy in Latin America. This chapter concludes
that unless significant advances in both socioeconomic development and the distribution of
economic resources occur, the likelihood of democratic consolidation in Latin America and the
Caribbean will be inhibited.
170
References
Almond, Gabriel A., and Sidney Verba. The Civic Culture: Political Attitudes and Democracy in
Five Nations. Princeton: Princeton University Press, 1963.
Armony, Ariel. The Dubious Link : Civic Engagement and Democratization. Palo Alto: Stanford
Banfield, Edward. The Moral Basis of a Backward Society. Chicago: The Free Press, 1958.
Boix, Carles. Democracy and Redistribution, Cambridge Studies in Comparative Politics.
Cambridge, UK ; New York: Cambridge University Press, 2003.
Campbell, Tim. The Quiet Revolution : Decentralization and the Rise of Political Participation
in Latin American Cities, Pitt Latin American Series. Pittsburgh, Pa.: University of
De Ferranti, David M. Inequality in Latin America : Breaking with History?, World Bank Latin
American and Caribbean Studies. Viewpoints. Washington, DC: The World Bank, 2004.
Fajnzylber, Pablo, Lederman Daniel, and Loayza Norman. "Inequality and Violent Crime."
Journal of Law and Economics 45, no. 1 (2002): 1.
Fajnzylber, Pablo, Daniel Lederman, and Norman Loayza. Determinants of Crime Rates in Latin
America and the World : An Empirical Assessment, World Bank Latin American and
Caribbean Studies. Viewpoints. Washington, D.C.: World Bank, 1998.
Finkel, Steven E, Anbal Prez-Lin, and Mitchell A. Seligson. "The Effects of U.S. Foreign
Assistance on Democracy Building, 1990-2003." World Politics 59 (2007): 404-39.
Fukuyama, Francis. Trust: The Social Virtues and the Creation of Prosperity. New York: The
Free Press, 1995.
Harrison, Lawrence E. Underdevelopment Is a State of Mind: The Latin American Case. Boston,
MA: Center for International Affairs, Harvard University and University Press of
America, 1985.
IADB. "Education, Science and Technology in Latin America and the Caribbean, a Statistical
Compendium of Indicators."
Inter-American Development Bank, 2006.
Inglehart, Ronald. Culture Shift in Advanced Industrial Society. Princeton: Princeton University
Press, 1990.
. "The Renaissance of Political Culture." American Political Science Review 82, no. 4
(1988): 1203-30.
. "Trust, Well-Being and Democracy." In Democracy and Trust, edited by Mark E.
Warren, 88-120. Cambridge: Cambridge University Press, 1999.
Inglehart, Ronald, and Christian Welzel. Modernization, Cultural Change, and Democracy : The
Human Development Sequence. Cambridge, UK ; New York: Cambridge University
Press, 2005.
Jackman, Robert W., and Ross A. Miller. "Social Capital and Politics." Annual Review of
Political Science 1, no. 1 (1998): 47-73.
Karl, Terry Lynn. "Economic Inequality and Democratic Instability." Journal of Democracy 11,
no. 1 (2000): 149.
171
Lederman, Daniel, Loayza Norman, and Menendez Ana Maria. "Violent Crime: Does Social
Capital Matter?" Economic Development and Cultural Change 50, no. 3 (2002): 509.
Lipset, Seymour Martin. Political Man: The Social Basis of Politics. 1981 (expanded edition) ed.
Baltimore, MD.: Johns Hopkins University Press, 1961.
. "Some Social Requisites of Democracy: Economic Development and Political
Legitimacy." American Political Science Review 53 (1959): 65-105.
Luke, Douglas A. Multilevel Modeling. Thousand Oaks, Calif.: Sage Publications, 2004.
Muller, Edward N., and Mitchell A. Seligson. "Civic Culture and Democracy: The Question of
the Causal Relationships." American Political Science Review 88 (1994): 635-54.
. "Insurgency and Inequality." American Political Science Review 81 (1987): 425-51.
Newton, Kenneth. "Trust, Social Capital, Civil Society, and Democracy." International Political
Science Review 22, no. 2 (2001): 201.
Norris, Pippa. The Democratic Phoenix: Reinventing Political Activism. Cambridge: Cambridge
Przeworski, Adam. Democracy and Development : Political Institutions and Well-Being in the
World, 1950-1990. Cambridge: Cambridge University Press, 2000.
Putnam, Robert. "The Strange Death of Civic America." The Independent 1996, 13.
Putnam, Robert D. Bowling Alone: The Collapse and Revival of American Community. New
York: Simon & Schuster, 2000.
. Making Democracy Work: Civic Traditions in Modern Italy. Princeton, NJ: Princeton
. "Tuning in, Tuning Out: The Strange Disappearance of Social Capital in America." PS,
Political Science & Politics 28, no. 4 (1995): 664.
Seligson, Mitchell A. "Can Social Capital Be Constructed? Decentralization and Social Capital
Formation in Latin America." In Developing Cultures : Essays on Cultural Change
edited by Lawrence E. Harrison and Jerome Kagan, xvi, 399 p. New York: Routledge, 2006.
. "The Renaissance of Political Culture or the Renaissance of Ecological Fallacy."
Comparative Politics 34 (2002): 273-92.
Theiss-Morse, Elizabeth, and John R. Hibbing. "Citizenship and Civic Engagement." Annual
Review of Political Science 8, no. 1 (2005): 227-49.
Uslaner, Eric M. The Moral Foundations of Trust. New York: Cambridge University Press, 2002.
. "Producing and Consuming Trust." Political Science Quarterly 115, no. 4 (2000): 569-
90.
. "Social Capital, Television, and The "Mean World": Trust, Optimism, and Civic
Participation." Political Psychology 19, no. 3 (1998): 441-67.
Uslaner, Eric, M. . "Trust and Social Bonds: Faith in Others and Policy Outcomes
Reconsidered*." Political Research Quarterly 57, no. 3 (2004): 501.
Uslaner, Eric M. , and Mitchell Brown. "Inequality, Trust, and Civic Engagement." American
Politics Research 33, no. 6 (2005): 868.
Verba, Sidney, Kay Lehman Schlozman, and Henry E. Brady. Voice and Equality : Civic
Voluntarism in American Politics. Cambridge, Mass.: Harvard University Press, 1995.
172
Statistical Analysis
Appendix VII.1
Dependent Variable : Civic participation in at least in one civic organization*

(Nineteen Latin American countries)
Odds Ratio Linearized Std. Err. t P>t [95% Conf.Interval]
Social Trust 1.001 0.000 2.850 0.004 1.000 1.002
Education 1.139 0.023 6.440 0.000 1.095 1.185
Approval current government 1.202 0.077 2.890 0.004 1.061 1.362
Age 1.006 0.001 5.930 0.000 1.004 1.008
Male 0.967 0.009 -3.820 0.000 0.950 0.984
Wealth 0.769 0.020 -10.230 0.000 0.732 0.809
Guatemala 0.715 0.080 -2.980 0.003 0.573 0.891
El Salvador 0.900 0.088 -1.080 0.282 0.742 1.091
Honduras 1.162 0.122 1.430 0.153 0.946 1.426
Nicargua 0.780 0.082 -2.360 0.018 0.635 0.959
Costa Rica 0.587 0.060 -5.250 0.000 0.481 0.716
Panama 0.512 0.062 -5.500 0.000 0.403 0.650
Colombia 0.904 0.100 -0.900 0.366 0.727 1.125
Ecuador 1.174 0.113 1.670 0.094 0.973 1.418
Bolivia 2.014 0.222 6.340 0.000 1.622 2.501
Peru 1.752 0.194 5.070 0.000 1.410 2.177
Paraguay 0.607 0.066 -4.600 0.000 0.491 0.751
Chile 0.698 0.080 -3.150 0.002 0.558 0.873
Uruguay 0.546 0.056 -5.890 0.000 0.447 0.668
Venezuela 0.959 0.118 -0.340 0.736 0.753 1.222
Dom. Rep. 1.527 0.157 4.120 0.000 1.248 1.868
Haiti 1.520 0.189 3.370 0.001 1.191 1.940
Jamaica 1.437 0.156 3.340 0.001 1.162 1.778
Guyana 1.394 0.150 3.090 0.002 1.129 1.720
*Logistic Regression Model; standard errors take into account design effect
Reference country: Mexico
Appendix VII.2
Variance Decomposition
Dependent Variable: Interpersonal Trust
Random-effects Parameters
Between groups .103
Within groups .927
Total 1.020
Percentage of variance explained by between groups differences: 10%
Statistical significance : 0.000; Number of observations: 32.274; Number of groups: 21
173
Appendix VII.3
Dependent Variable : Support for Democracy (ING4)*

(Eighteen Latin American countries)
Coef. Linearized Std. Err. t P>t [95% Conf. Interval]
Social Trust 0.049 0.007 7.110 0.000 0.035 0.062
Religious org. 1.406 0.494 2.850 0.004 0.438 2.375
Parents association -3.315 0.772 -4.290 0.000 -4.830 -1.801
Community org. -0.439 0.806 -0.540 0.586 -2.019 1.142
Professionals, etc. org. -2.373 1.096 -2.170 0.031 -4.523 -0.223
Political party or movement -2.401 1.204 -1.990 0.046 -4.763 -0.039
Education 2.871 0.298 9.620 0.000 2.286 3.456
Approval current government 10.920 1.069 10.210 0.000 8.823 13.018
Satisfaction with democracy 0.135 0.010 13.600 0.000 0.115 0.154
Economic situation -0.643 0.242 -2.650 0.008 -1.118 -0.167
Wealth 0.428 0.128 3.350 0.001 0.177 0.678
Age 0.140 0.013 10.910 0.000 0.115 0.166
Male 0.626 0.352 1.780 0.075 -0.064 1.316
Guatemala 0.410 1.525 0.270 0.788 -2.582 3.402
El Salvador 5.378 1.865 2.880 0.004 1.719 9.036
Honduras -5.943 1.487 -4.000 0.000 -8.859 -3.027
Nicaragua 0.410 1.525 0.270 0.788 -2.582 3.402
Costa Rica -4.407 1.575 -2.800 0.005 -7.497 -1.317
Panama 7.635 1.443 5.290 0.000 4.805 10.466
Colombia -13.249 1.968 -6.730 0.000 -17.109 -9.389
Ecuador 1.335 1.503 0.890 0.375 -1.612 4.282
Bolivia -0.286 1.623 -0.180 0.860 -3.470 2.898
Peru -2.343 1.425 -1.640 0.100 -5.139 0.453
Chile -6.713 1.696 -3.960 0.000 -10.041 -3.386
Uruguay 3.615 1.730 2.090 0.037 0.221 7.009
Venezuela 8.650 1.550 5.580 0.000 5.610 11.690
Dom. Rep. 8.462 2.245 3.770 0.000 4.058 12.865
Haiti 8.672 1.503 5.770 0.000 5.724 11.621
Jamaica 8.961 2.192 4.090 0.000 4.662 13.260
Guyana 7.828 1.577 4.960 0.000 4.735 10.920
Constant 2.914 1.606 1.810 0.070 -0.237 6.064
*Linear Regression Model (OLS); standard errors take into account design effect
Number of observations: : 25.468; R-squared: 0.11
174
Appendix VII.4
Dependent Variable : Political Tolerance*

(Nineteen Latin American countries)
Coef. Linearized Std. Err. t P>t [95% Conf. Interval]
Social Trust 0.038 0.007 5.460 0.000 0.024 0.051
Religious org. -0.876 0.475 -1.840 0.065 -1.807 0.056
Parents association -2.335 0.711 -3.290 0.001 -3.729 -0.941
Community org. -1.004 0.793 -1.270 0.205 -2.559 0.551
Professionals, etc. org. 1.656 1.058 1.570 0.118 -0.419 3.731
Political party or movement 4.175 1.082 3.860 0.000 2.053 6.296
Approval current government -8.280 0.892 -9.290 0.000 -10.029 -6.532
Education 3.899 0.301 12.970 0.000 3.309 4.488
Wealth 0.330 0.123 2.680 0.007 0.089 0.571
Edad 0.006 0.013 0.460 0.648 -0.019 0.031
Male 2.431 0.329 7.380 0.000 1.785 3.077
Guatemala -2.566 1.354 -1.890 0.058 -5.222 0.091
El Salvador 0.350 1.274 0.270 0.784 -2.149 2.848
Honduras -8.805 1.592 -5.530 0.000 -11.926 -5.683
Nicaragua -3.015 1.679 -1.800 0.073 -6.308 0.278
Costa Rica 4.016 1.484 2.710 0.007 1.104 6.927
Panama -7.471 1.800 -4.150 0.000 -11.001 -3.941
Colombia -3.708 1.491 -2.490 0.013 -6.632 -0.783
Ecuador -11.038 1.392 -7.930 0.000 -13.769 -8.307
Bolivia -11.977 1.375 -8.710 0.000 -14.673 -9.281
Peru -4.242 1.466 -2.890 0.004 -7.117 -1.367
Paraguay 0.704 1.635 0.430 0.667 -2.502 3.910
Chile -1.003 2.065 -0.490 0.627 -5.053 3.046
Dom. Rep. 5.469 1.914 2.860 0.004 1.716 9.223
Haiti 8.920 1.856 4.810 0.000 5.280 12.560
Jamaica 3.798 1.668 2.280 0.023 0.526 7.069
Guyana 7.690 1.998 3.850 0.000 3.772 11.608
Constant 15.711 1.569 10.010 0.000 12.633 18.788
*Linear Regression Model (OLS); standard errors take into account design effect
R-squared: 0.09
175
VIII. National Identity and Ethnic Minorities in the

Americas
Daniel Moreno
Abstract
This chapter focuses on the strength of the bond between citizens and the national political
community in the sample of countries included in the AmericasBarometer 2006-7 by LAPOP.
Two research questions are explored: First, how does the ethnic minority status affect the sense
of national belonging among citizens? And second, what is the effect that the level of
development has on the sense of national belonging? A multi-level analysis is employed to
determine the effect that ethnic identity, as an individual level variable, and the level of
development as a country level variable have on the sense of national belonging. Evidence
suggests that, excepting three countries, ethnic minority status does not have a relevant effect
on the strength by which citizens relate to the national political community; this finding seems
to be independent from the level of development of a country. At the national level, human
development appears to be an important predictor of the average strength of the bond between
citizens and the nation.
This chapter focuses on the connection citizens have to their national political
community. In particular, it explores the relationship between ethnic identity and the sense of
national belonging, taking into consideration the fact that it is possible for there to be important
differences in this relation according to each countrys level of development.
This topic has become more relevant in the last two decades, during which ethnic
identities have become more important as part of a trend that could very well be called global
(Brysk 2000; Connor, 1994). In the 1990s, the breakdown of Yugoslavia and the civil war in
Rwanda were the most visible signs, albeit not the only ones, of the potentially negative
consequences of the exacerbation of identity on political stability in those countries. In present
times, there are citizens in different Latin American countries who question the legitimacy of
the national state itself; this makes observers point at the bonds between citizens and the State
as a possible explanation for differences in the quality of democracy in the region.
In this chapter responses are sought for the following questions: Following a world-wide
trend of politicization of ethnic identities, has this trend become a threat to democracy in the
This chapter has been prepared thanks to financial support from USAID/Bolivia and the Center For the Americas
at Vanderbilt University. Suggestions and commentary by Mitchell A. Seligson and the LAPOP team members
were particularly useful for this paper. Any errors and omissions are only the authors own.
Ph.D. candidate at Vanderbilt University; Fellow of the Center for the Americas 2006-2008.
177
Americas? What role does the level of national development play in the relation between
identities and the feeling of national belonging? What other variables may explain the relative
strength of the link between the citizen and the national political community? In order to
perform this analysis, the entire sampling of the AmericasBarometer for the year 2006-07 is
used.
The National Political Community
The sense of national belonging is one of the most important conditions for democracy.
In order to feel motivated to participate in democracy as well as to abide by the law, citizens
need to recognize the State they live in as legitimate. From this insight, and following the line
drawn by Mill (1993), Rustow suggests that the strength of the bond between citizen and State
is an indispensable condition for democracy (1970).
Being part of a nation in some way implies that ones own destiny is joined to that of the
rest of its citizens. It also implies that all citizens accept and recognize the legitimate power of
the State. The national political community, the nation of citizens, can therefore be understood
as an imaginary community, as it has been defined in Andersons (1993) seminal work; that is,
as a community of persons who, without knowing each other, imagine they share the same bond
of fraternity and equality.
In spite of the importance of this area of study, comparative politics has paid very little
attention to this matter, and empirical studies based on quantitative information have been
scarce or nonexistent (Juviler and Stroschein 1999). A notable exception has been the Smith and
Jarkko study (2001).
Most studies that deal with national identity do so only tangentially, often stemming
from the greater concept of system support, as part of what has been called diffuse support
(Dalton 1999; Dalton 2004; Easton 1965; Easton 1975; Muller, Jukam and Seligson 1982;
Norris 1999). In this chapter, an explicit distinction is made between the concepts of support for
the system and that of belonging to a national political community. Although it seems clear that
feeling oneself part of a political nation is a necessary condition for showing support for the
political and institutional system, it should not be considered part of the same theoretical
construct.
National Belonging: Is it Necessary for Democracy?
To what extent is this sense of belonging to a national political community important for
democracy? To respond to this question, two elements of that support that are central to
democracy will be analyzed: system supportwhich refers to the relation between citizens and
State, as well as political tolerancewhich speaks of a horizontal relationship between citizens
within the framework of a democratic State.
178
The LAPOP 2006-2007 data is revealing and demonstrates that the feeling of belonging
to a national political community has a positive effect on these democratic values. Figure VIII.1
shows the relationship of national pride, described and analyzed below, with support for the
political system, measured as described in chapter IV of this volume.
60
50
System support
40
30
20
0 17 33 50 67 83 100
Pride in nationality
Source: AmericasBarometer 2006, by LAPOP
Figure VIII.1: Effects of pride in nationality on support for the system
This figure indicates that the correlation is clear: people who feel greater pride in their
nationality tend to be more supportive of the countrys political system. The relationship is
strong (r=0.230) and statistically significant (p<0.001). Regarding support for the system and
national pride, it is possible, however, that both variables measure concepts that are too closely
connected to establish a causal relationship between the two variables.
To determine the existence of a causal relationship, another value that is central to

democracy, political tolerance, is examined. Tolerance has to do with citizen respect for the
right of others to think and act in a manner that is different and perhaps opposite from their own
(Adorno, et al. 1950; Gibson 1992). This variable is described and analyzed in chapter IX of
this volume. Figure VIII.2 shows the relationship between political tolerance and the measure of
acceptance with regards to the idea of common values, the other variable considered in this
chapter.
179
65
60
Political tolerance
55
50
45
0 17 33 50 67 83 100
Shared Values
Figure VIII.2: Effects of the belief in common values on political tolerance
The correlation between the two variables is strong and significant (r=0.18, p<0.001).
The appropriate test of robustness of this relationship is a lineal regression analysis in which
statistical control factors that generally explain political tolerance (such as education, age,
income and gender) are taken into account. Each additional ten points on the common values
scale increase the average political tolerance by 2 units on its 0 to 100 scale. Hence, shared
values matter for this vital element of democratic political culture.
This relationship suggests furthermore the possibility of assuming a causal relationship

by means of which a greater sense of attachment to the national political community, measured
by the scale of common values, results in a greater level of political tolerance. This hypothesis
also has its own logical consistency; the more a person feels part of a national community, the
greater the persons predisposition to showing respect for different members of this community
will be. In this sense, to feel part of the national political community could lead to the
recognition of the other as a citizen, or as the legitimate other as Mouffe (2000) puts it.
The Strength of the National Political Community in the Countries Studied
This chapter empirically analyzes two variables that refer to the strength of the bonds
that link citizens to the national political community. The first stems from the following
question used in the LAPOP survey: How proud are you of being (Mexican, Guatemalan, etc.)?
180
The second variable comes from the question that reads In spite of our differences, as
(Mexicans, Guatemalans, etc.) we have values that unite us as a country. To what extent do you
agree with this statement? Both questions were initially measured on a 1 to 7 scale, in which 1
means Not at all and 7 Very much; values were recoded into a zero to 100 scale to present
the results more easily. These questions are normally used in studies regarding this topic, and
are supported in the literature (Norris 1999).
The two variables measure two different dimensions of the bond between individuals
and the national political community. The first directly addresses the level of pride that a person
feels for his or her nationality and is based on the idea that the feeling of belonging to a national
community is reflected in the pride of being a part of it. The second focuses on the idea that the
nation exists beyond the territory of the State, in the values shared by its citizens (or at least in
the belief that these values exist). These two variables have a correlation of .247 (p<0.001),
which means that they are related but not very high, which suggests that the variables do indeed
measure different dimensions of the concept.
Figure VIII.3 presents national averages for the two variables focused on in this chapter.
Shared Values
100 Pride in nationality
97.4 97.0 96.2 95.1

93.9 93.5 93.4 93.3 93.2 92.9 92.9
92.3 91.5
88.8 87.7 87.3
80 86.4 86.4 86.2
83.0 80.7 85.4 84.5 86.3 86.9 86.7 85.3 84.9
78.2 78.9 83.9 82.9
77.8 76.9 78.2 79.0
80.9
74.8 75.9
National mean
77.7
71.3 70.8 70.8
60
60.9
40
20
0
Mexico
Ecuador
Canada
Jamaica
Guyana
Nicaragua
Panama
Peru
Honduras
Paraguay
Uruguay
Venezuela
Guatemala
USA
Haiti
Chile
Colombia
Brazil
El Salvador
Bolivia
Costa Rica
Dominican Rep.
Error bars: 95% CI
Figure VIII.3: Averages for National pride and the belief in common values, per country
It can be seen that national pride levels are relatively high in all countries in the study,
although the Is corresponding to the error bars suggest that countries can be classified into
181
four groups1. The highest averages belong to the Dominican Republic, Costa Rica, and
Venezuela. The next group of countries consists of Ecuador, Mexico, El Salvador, Paraguay,
Colombia, Honduras, the United States, Uruguay, Canada and Jamaica. The third group is made
up of Chile, Peru, Guyana, Bolivia, Brazil, Nicaragua, Guatemala and Panama. The lowest
registered value belongs to Haiti.
The variation is much higher for the item that measures the level of agreement with the
idea of common values. The United States is the country with the highest national average for
this variable, and seems to stand out among the rest. Following the U.S. is the group including
the Dominican Republic, Costa Rica, Jamaica, Canada, Venezuela, Uruguay, and Mexico. Next
is the group is made up of Colombia, Guyana, Honduras, Ecuador, Chile, El Salvador,
Paraguay, Brazil and Guatemala. Much lower values belong to Bolivia and Nicaragua, whereas
Panama has an even lower average which sets it apart from any other country in the series.
Previous analysis suggests that national pride tends to be higher in stable and
institutionalized democracies than in countries with more fragile democracies (Smith and
Jarkko 2001). Although the data shows that countries with more stable democracies have higher
levels of national pride, the relationship does not appear to be linear: the most consolidated
democracies in the region, Canada and the United States, show lower averages than those of
less consolidated democracies such as Ecuador or the Dominican Republic. This explanation,
therefore, does not seem to clearly explain the opinions expressed in countries in the region.
An alternative explanation for these differences is the level of development in each

country. Following, I present the national averages of each country according to the level of
human development as measured by the United Nations Development Program2 (UNDP 2006)
Figures VIII.4 and VIII.5 show the relation between the national averages of the two focus
variables and the Human Development Index calculated for each country.
1
The I that each bar includes represents the confidence interval of the average presented. As a rule of thumb, if
the I s of two bars overlap, it cannot be assumed that the averages are statistically different.
2
The Human Development Index is an indicator composed of economic (GNP per-capita calculated according to
Purchasing Power Parity), educational (the countrys educational index), and health (the populations life
expectancy) aspects. The UNDP calculates the value of the index for every country in the world bi-annually. For
further information see http://hdi.undp.org.
182
100
Dominican Rep.
Costa Rica
Venezuela
Ecuador
95
Mexico
El Salvador Paraguay
Uruguay USA
Honduras Colombia
Jamaica Canada
90
Chile
Peru
Bolivia
Guyana
Brazil
Nicaragua
85
Guatemala
Panama
Haiti
80
0.40 0.50 0.60 0.70 0.80 0.90 1.00

Source: AmericasBarometer 2006, by LAPOP and UNDP 2006
Figure VIII.4: National Pride Averages by HDI
It is clear that the relation between the level of human development and national pride is
positive. Excluding Panama from the sample, the correlation is strong and significant (r=0.535,
p<0.05). In spite of this strength, the relationship is not unequivocal nor is it monotonic. The
further a country is from the line of regression, the lower the predictability of its average level
of pride according to its level of human development. For example, Costa Rica, Brazil,
Dominican Republic and Ecuador are above the expected average value, whereas Brazil, Chile
and Guatemala are below the expected levels. Panama was excluded from the sample because it
is clearly an outlier, as will be evident below. There are many factors that might explain the
pride citizens have in their nationality, ranging from the performance of the national team in
world soccer matches, which could be the case in Ecuador (Seligson, et al. 2006b) to the
internal perception of the countrys image abroad.
As Graph VIII.5 below shows, the relationship between human development and
common values does not seem to be very strong at first look; this can be explained again by
the presence of Panama. This is an atypical case, an outlier, which pulls the regression line;
once this case is removed from the sample, the relation becomes strong and significant
(r=0.497, p<0.05).
183
90 USA
Dominican Rep. Costa Rica
Jamaica Venezuela Canada

Uruguay
Mexico
Colombia
80 Honduras Guyana
Ecuador Chile
Haiti
El Salvador
Guatemala Paraguay
Brazil
Nicaragua Peru
70
Bolivia
Panama
60
0.40 0.50 0.60 0.70 0.80 0.90 1.00

Source: AmericasBarometer 2006, by LAPOP and UNDP
Figure VIII.5: Average according to the idea of common values and HDI
Considering national averages at the country level, the Human Development Index
appears to be a factor that weighs heavily when predicting both national pride and agreement
with the idea of common values. Using national averages in this type of analysis is helpful when
comparing countries or registering the effects of contextual variables on matters germane to this
study. However, national averages do not tell the whole story; the great differences among
citizens within each country are best explained by individual characteristics, and not those of
the country. This problem in the analysis is known as the ecological fallacy and is a common
failing in the analysis of cross-national survey data (Seligson 2002).
National Belonging and Ethnic Identity. Questioning Assumptions
One of the central defining characteristics of individuals, which in theory, at least, can
become a focus point for strengthening the bond between the citizen and the state or nation, is
ethnic identity. There are authors for whom ethnic diversity represents an obstacle to liberal
democracy (Chua 2003; Horowitz 1985; Rabushka and Shepsle 1972; Snyder 2000); the main
argument for this point of view is that ethnic identities, understood as the primordial
belonging to a group (Geertz 1963; Stack 1986; Van Evera 2001), create stronger alliances
among members than those that can be created by national states that are made up of different
ethnic groups. Tragic examples such as Yugoslavia and Rwanda in the 1990s are commonly
184
used as evidence for this supposed contradiction between democracy and strengthened ethnic
identity.
However, this position has increasingly been debated, principally using a less
essentialist conception of ethnic identity, in which identities are understood as complex, fluid,
and malleable social constructs. This means that for some authors, ethnic diversity is not
necessarily a problem for the democratic stability of a country (Abizadeh 2002; Chandra 2006;
Gutmann 2003) and that ethnic differences are relevant only when they go hand in hand with
social and economic differences that systematically transform ethnic patterns into mechanisms
for social stratification.
One implication of the primordialist position in this debate is that ethnic identities
would tend to be stronger than national identity. This means that ethnic minorities in the
countries studied may exhibit a weaker feeling of belonging to the national community than
individuals who belong to the majority ethnic group. In order to test this theory, a dichotomous
variable was created assigning a value of 1 to persons who identified themselves as part of a
minority group in the country and 0 to those who identified themselves as part of the majority
ethnic group.
The variable was generated by the question: How do you describe yourself? This
question was included as an item in the LAPOP questionnaire with slight variations in question
wording and adjustments to the possible categories listed for each country. In most Latin
American countries the majority was considered to be a group constituted by a combination of
respondents who identified themselves as white and mestizo (mixed Native American and
European), whereas indigenous (Native American) and afro descendants are part of the
minority category;3 in the United States those who identified themselves as racially white
were classified the majority, whereas in Canada the individuals classified as part of the majority
did not identify themselves with any other category except Canadian (principally French-
Canadian). In Brazil, individuals who identified as branco and pardo where coded as non-
minorities. In Haiti, those identified as white were classified as the minority. Three of the
twenty-two countries surveyed by LAPOP in 2006-2007 were not included in the analysis:
Paraguay, where there were no questions regarding ethnic identification on the questionnaire, as
well as Jamaica and Guyana, where classifying individuals according to the options on the
questionnaire is particularly complicated.4
Figure VIII.6 illustrates the percentage of people in each country who were coded as
belonging to an ethnic minority. It can be seen that the relative number of people who were
3
Although it is true that in some Latin American countries such as Guatemala and Bolivia the majority population
could be indigenous (Gurr 1993; Van Cott 2005; Yashar 2005), the national governments have historically been
made up of citizens who can be classified as mestizos or white. Of course, the recent election of Evo Morales
in Bolivia represents a substantial change in these power relations, and this has had important effects on citizen
opinions and attitudes vis--vis the State, as confirmed by LAPOP studies (Seligson, et al. 2006). Despite their
relative majority, and following the rule of thumb applicable for other Latin American countries, indigenous groups
in Bolivia and Guatemala can be classified as minorities in terms of their access to power.
4
Details of those questionnaires applied in the surveys in each of the countries are available on the LAPOP internet
web page (www.lapopsurveys.org).
185
classified as part of the minority varies greatly from country to country, from over 40% in
Guatemala to less than 5% in Nicaragua.
100% Ethnic Minority Status

Minority
Non-minority
80%
60%
40%
20%
0%
Mexico
Ecuador
Canada
Panama
Nicaragua
Peru
Honduras
Uruguay
Guatemala
Venezuela
USA
Haiti
Chile
Colombia
Brazil
Bolivia
El Salvador
Costa Rica
Dominican Rep.
Figure VIII.6: Per-country ratio of persons according to ethnic

minority status.
What effect does being part of an ethnic majority or minority have on the individuals
sense of belonging to the political community? After running a linear regression analysis for
each of the two dependent variables that include a persons age, sex, educational level and
income as statistical controls, the differences between the ethnic majority and minority are
evident in only a handful of countries. Figure VIII.7 shows the effect of the differences between
majority and minority groups in the countries studied, where the difference is robust after
introducing the social demographic controls. The dependent variable is pride in ones own
nationality.
186
100
Ethnic majority
Ethnic minority
80
National pride
60
95 93
89 88
86 84
40 82
79
20
0
Canada Honduras Peru Panama

Figure VIII.7: Pride in Nationality by Majority Status
(countries with significant differences)
Only in Canada, Honduras, Peru and Panama do minorities feel less proud of their
nationality than the ethnic majority. In Canada this factor reflects the greatest differences, as the
figure shows. It is notable that in the few cases where there are differences between the majority
and the ethnic minority, the minority groups averages are lower than the majoritys. Therefore,
only in these countries is the result what could be expected from a primordialist vision of
ethnic identities.
When comparing the variable that measures the level of agreement with the idea of
common values it can be seen that the results are very similar. Figure VII.8 shows this, as
follows:
187
100
Ethnic majority
Ethnic minority
Agreement with idea of common values

80
60
87
40 79 79
72 72
66
64
60
20
0
Canada Honduras Peru Panama
Source: AmericasBarometer by LAPOP, 2006

Figure VIII.8: Belief in Common Values by Majority Status
(Countries with significant differences)
The countries in which ethnic identity has some effect on the sense of national
belonging are the same one appearing in the previous case. Again, the greatest differences are in
Canada. The main difference in evidence upon comparing these results with those viewed
previously, is that people who identify themselves as part of ethnic minority groups in Panama
remain more convinced that there are common values that bond Panamanians together than the
members of the white/mestizo majority. The minority variable in Panama shows a paradoxical
behaviour: one of the variables is affected positively, the other is affected negatively.
Aside from the paradoxical case of Panama, it can be seen that the countries in which
the differences between majority and minority groups are most important are Canada, Honduras
and Peru. Those who identify themselves as indigenous, black or mulatto in Honduras and Peru,
and as French-Canadian or members of another racial group in Canada, systematically exhibit a
lesser attachment to the national political community of their respective countries than those
who feel they are part of the ethnic group coded as the majority. This suggests a weaker
inclusion of these populations in the national project represented by these three states.
The key question that arises from these observations, then, is: What factors cause ethnic
identity to have a significant effect on the sense of belonging to the national political
community in one country and not in others? Why are ethnic differences on these survey
188
questions relevant in Peru but not in Ecuador, relevant in Honduras but not in Nicaragua? In
other words, what do Canada, Honduras and Peru have in common that differentiates them from
the rest of the countries in the group?
Combining Individual Factors and Contextual Factors
There are statistical tools that allow us to determine the influence that national factors,
such as the Human Development Index or the Gross National Product, exert simultaneously
with the effect that individual factors have, such as the persons gender and age, or personal
status as a member of an ethnic majority or minority. One of these tools is multi-level statistical
analysis, or hierarchical linear modelling, which allows for the combining of information
regarding national and individual factors (Luke 2004; Raudenbush and Bryk 2002). A multi-
level analysis model is basically a generalized linear regression model (GLM) in which the
independent variable effects are allowed to vary randomly, that is to say, that they are not
constant.5
Using the combined base of countries studied by LAPOP in 2006-20076, a series of

statistical analyses were carried out; various equations were used to test the relations between
both dependent variables and independent variables at the individual level (gender, age, ethnic
minority status), as well as the Human Development Index for the year 2006 as a variable
corresponding to the country.
In the same way, an interaction term between an individual level variable (ethnic status)
and a national level (HDI) was included7. If this term proved to be statistically significant, we
would then be able to assume that the effect of being part of an ethnic minority varies according
to the level of human development in these countries. Table VIII.1 which we present next,
presents the most important results for each variable8.
5
The model used is known as Random Effects Maximum Likelihood (REML).
6
The combined base excludes Guyana, Jamaica and Paraguay for lack of information in the minority variable.
Panama is excluded as well due to its paradoxical behaviour regarding both variables. Therefore, 15 countries with
a total of 24,426 interviews are considered here.
7
Given that the number of countries is relatively small, the inclusion of more than one variable at this level would
be a problem.
8
The results presented for all variables come from models with random intercepts and slopes estimated for the
variable minority in which the interaction term was not included in order to avoid collinearity. The interaction
term was incorporated into subsequent models based on those now described. Results in the shaded cells of table
VIII.1 correspond to the models including the interaction term.
189
Table VIII.1: Results of multi-level analysis on the two dependent variables

Dependent Var. National Pride+ Common Values+
Independent Var.
Minority status -0.64 (-0.85) -1.60 (-2.37)*
Age 0.36 (4.18)** 0.50 (4.95)***
Educational Level -0.06 (-0.37) 1.80 (8.87)***
Gender (Women) -0.11 (-0.46) 0.57 (1.99)*
HDI 24.9 (2.69)** 24.87 (2.20)*
HDI X Minority -11.81 (-1.54) -10.78 (-1.51)
Constant 70.7 (9.83)*** 54.49 (6.19)***
SD Minority/ Constant / 2.6 / 4.1 / 20.3 1.88 / 4.99 / 23.6
Residual++
N individuals / country 27,179 / 18 26,684 / 18
+ Values are coefficients (z value)
++ Random Effects
* p<.05; ** p<.01; *** p<.001
In the combined eighteen-country database that is considered here, being part of an

ethnic minority significantly diminishes the average belief in common values, but has no effect
on the variable that measures national pride. This is the average result for the entire region (it
has been observed that averages can be different in each particular country), and this is
independent from the level of development in the country, the persons age, educational level
and gender.
Age has a highly significant positive effect on the sense of belonging to a national
political community. Once age is controlled by other factors, the older people tend to feel more
pride in their nationality and to be surer of the existence of common values than do younger
citizens. Therefore, young people feel less attached to the nation that they live in, perhaps
strongly marked by the idea of universal citizenship that comes out of the information
technology revolution.
The other two variables at the individual level which are included in the model are
educational level and gender. Both affect significantly only the idea of common values. The
higher a persons level of education, the higher tends to be her conviction that there are values
which unite the citizens in her country. Women are found to agree more with the existence of
common values in their countries than men; this might indicate that men see internal differences
in each country as something more unsolvable than women do.
The Human Development Index, a country level variable included in this analysis, has a
positive and significant effect on national pride as well as on the variable of common values, as
mentioned before. The higher the level of human development in a country, the greater the
citizens pride in their nationality and the more they tend to believe in the existence of values
that unite the country. This result is independent of peoples educational level and the other
factors considered in the model.
190
In the analysis, the variable that has no statistical significance is the interaction term
between human development and minority status. This means that, among countries under
consideration, the level of development does not mediate the effect of minority status on the
levels of feeling of belonging to the national political community. However, as previously
stated, the number of countries available for analysis is small and represents a truncated
distribution of the universe of countries in the world. It is possible, with a world-wide sample,
that the effect of belonging to an ethnic minority on the individuals connection with the
national political community might vary according to the countrys level of development
although probably not linearly.
In short, it can be said that in the sample made up of the Americas, ethnic minority or
majority status has a small or nonexistent effect on the strength of the bond between individuals
and the national political community. People who are classified here as members of an ethnic
minority are less persuaded of the existence of common values among citizens in their countries
than those who belong to the group characterized as majority, yet their level of national pride is
indistinct from that of the majority. These findings suggest that some research in this field may
have exaggerated the impact of ethnic differences for variables related to the bond with the
national political community.
Conclusions
Three main conclusions can be drawn from the analysis presented in this chapter. First,
ethnic minority status in and of itself has no effect on the variables that assess the strength of
the bond between the citizen and the national political community. The heightened focus on
ethnic identities in recent years seems to pose no risk to consolidated nations or their
democracies. National pride is, generally, the same for ethnic minorities and majorities, whereas
the idea of the existence of common values is slightly lower among minorities than among
national ethnic majorities. There are, however, three countries in which minorities appear to be
less integrated into the national political community: Canada, Honduras and Peru. These three
cases merit further consideration in order to uncover factors that are common to the three
nations or specific to each country, which might explain the potentially problematic insertion of
minorities into the national political community.
Second, a countrys human development has an effect on national pride. It has been
observed that the higher the level of human development, the more citizens feel proud of their
nationality and the greater their belief that they are brought together by shared values. However,
contrary to expectations, the level of development does not seem to affect the relationship
between ethnic minority status and national belonging. It has been found that ethnic differences
matter in countries that are in very different places in the UNDPs international human
development ranking: Canada is number 6, Peru number 92 and Honduras number 117 among
the 177 countries for which the UNDP has information (UNDP 2006). It is probable that if the
level of development mediates the effect between ethnic minority status and the sense of
belonging to a national political community, it does so in a non-linear manner. In any case,
191
Limitations of this study derived from the use of a sample restricted only to the region make the
arrival to further reaching conclusions on this issue somehow problematic. 9
The third conclusion is that the variables that measure the bond between citizens and the
national political community can aid in understanding the possibilities of the state of democracy
in different countries. It has been shown that national pride is strongly related to support for the
system and that the idea of common values are strongly associated with political tolerance. In
the future, those interested in promoting democracy would do well to pay more attention to the
conditions in which citizens form part of the national political community as an important factor
in explaining the growth in the political culture of democracy.
Finally, it is necessary to point out the relevance that the study of the relationship
between ethnic identities and political participation has beyond the issue of inclusion in the
national political community considered in this chapter. While it has been shown here that
ethnic minority status matters for the strength of the political community in some countries, it is
still unclear how and under what conditions identities are relevant for more specifically political
behaviours and attitudes, such as voting preferences and ideological positioning. These are
questions that remain to be answered, and that the construction of datasets such as the
AmericasBarometer can help to clarify.
9
It is certainly possible that there are other factors that may explain the differences between these countries and
others in the region, which were neither taken into account theoretically nor empirically in this chapter. Finally, it
is not possible to discard the hypothesis which says that these differences are significant in each country for
contextual reasons that are specific to each country and not due to the existence of a common explanatory factor. In
any case, the only way to determine this is by means of a multi-level statistical analysis, using more cases at the
country level, which could incorporate more country-specific variables and in this way test the effect of other
factors on the variables we have focused on in this chapter.
192
References
Abizadeh, Arash. "Does Liberal Democracy Presuppose a Cultural Nation? Four Arguments."
American Political Science Review 96, no. 3 (2002): 495-509.
Adorno, TW, Frenkel-Brunswik Else, Daniel Levinson, and Nevitt Sanford. The authoritarian
personality. New York: Harper & Row publishers, 1950.
Anderson, Benedict. Comunidades Imaginadas. Mxico: Fondo de Cultura Econmica, 1993.
Brysk, Alison. From Tribal Village to Global Village: Indian Rights and International
Relations in Latin America. Stanford: Stanford University Press, 2000.
Chandra, Kanchan. "What is Ethnic Identity and Does It Matter?" Annual Review of Political
Science 9 (2006): 397-424.
Chua, Amy. World on fire. New York: Doubleday, 2003.
Connor, Walker. Ethnonationalism: The Quest for Understanding. Princeton: Princeton
Dalton, Russell. "Political Support in Advanced Industrial Democracies." In Critical Citizens,
edited by Pippa Norris. Oxford: Oxford University Press, 1999.
. Democratic Challenges, Democratic Choices: The Erosion of Political Support in
Advanced Industrial Democracies. Oxford: Oxford University Press, 2004.
Easton, David. A Systems Analysis of Political Life. New York: Willey, 1965.
. "A Reassessment of the Concept of Political Support." British Journal of Political
Science 5 (1975): 435-57.
Geertz, Clifford. "The Integrative Revolution: Primordial Sentiments and Civil Politics in New
States." In Old Societies and New States: The Quest for Modernity in Asia and Africa,
edited by Clifford Geertz. New York: Free Press of Glencoe, 1963.
Gibson, James. "The Political Consequences of Intolerance: Cultural Conformity and Political
Freedom." American Political Science Review 86, no. 2 (1992): 338-56.
Gurr, Ted Robert. Mnorities at Risk. A Global View of Ethnopolitical Conflict. Washington DC:
United States Institute for Peace Press, 1993.
Gutmann, Amy. Identity in democracy. Princeton - Oxford: Princeton University Press, 2003.
Horowitz, Donald. Ethnic Groups in Conflict. Berkeley: University of California Press, 1985.
Juviler, Peter, and Sherrill Stroschein. "Missing boundaries of comparison: The political
community." Political Science Quarterly 114, no. 3 (1999): 435-53.
Luke, Douglas. Multilevel Modeling: Quantitative Applications in the Social Sciences. Newbury
Park, CA: Sage, 2004.
Mill, John Stuart. Utilitarianism, On Liberty, Considerations on Representative Government,
Remarks on Bentham's philosophy. London- Vermont: Everyman, 1993.
Mouffe, Chantal. The democratic paradox. London: Verso, 2000.
Muller, Edward, Thomas Jukam, and Mitchell Seligson. "Diffuse Political Support and
Antisystemic Political Behavior." American Journal of Political Science 26, no. 2
(1982).
Norris, Pippa, ed. Critical Citizens: Global Support for Democratic Governance. Oxford:
Oxford University Press, 1999.
Rabushka, Alvin, and Kenneth Shepsle. Politics in Plural Societies: A Theory in Democratic
Instability. Columbia: C. Merrill, 1972.
193
Raudenbush, S. W., and A.S. Bryk. Hierarchical Linear Models: Applications and Data
Analysis Methods. 2nd ed. Newbury Park: Sage, 2002.
Rustow, Dankwart. "Transitions to Democracy: Toward a Dynamic Model." Comparative
Politics 2, no. 3 (1970): 337-63.
Seligson, Mitchell A. "The renaissance of political culture or the renaissance of the ecological
fallacy?" Comparative Politics 34, no. 3 (2002): 237-92.
Seligson, Mitchell, et al. Auditora de la democracia. Informe Bolivia 2006. Cochabamba:
USAID - LAPOP - CIUDADANIA, 2006a.
. Auditoria de la democracia. Ecuador 2006. Quito: CEDATOS - LAPOP, 2006b.
Smith, Tom, and Lars Jarkko. National Pride in Cross-National Perspective National Opinion
Research Center, 2001 [cited. Available from
http://www.issp.org/Documents/natpride.doc.
Snyder, Jack. From Voting to Violence: Democratization and Nationalist Conflict. New York:
Norton, 2000.
Stack, John, ed. The Primordial Challenge: Ethnicity in the Contemporary World. New York:
Greenwood Press, 1986.
UNDP. Human Development Report 2006 2006 [cited. Available from http://hdr.undp.org.
Van Cott, Donna Lee. From movements to parties in Latin America: The evolution of ethnic
politics. New York: Cambridge University Press, 2005.
Van Evera, Stephen. "Primordialism Lives!" APSA - CP: Newsletter of the organized section in
comparative politics of the American Political Science Association 12 (2001): 20-22.
Yashar, Deborah. Contesting Citizenship in Latin America: The Rise of Indigenous Movements
and the Postliberal Challenge. Cambridge ; New York: Cambridge University Press,
2005.
194
IX. Challenges of Tolerance in the Americas1

Diana Orcs
Abstract
This chapter analyzes the importance of political and social tolerance for democracy. Citizens
who have higher levels of tolerance are more willing to accept and respect the political rights
and civil liberties of the minorities and those whom they may disagree. If the rights of minorities
to express themselves freely are not guaranteed, how can democracy be sustained over the long
term? For this reason tolerance is of great importance for the persistence of democracy as a
form of government. In this chapter, a comparison of political and social tolerance is made
across countries in which differences between political and social tolerance are partly attributed
to cultural factors such as British, French, and Spanish colonial traditions. Nonetheless, the
primary task of this research is to explain which individual-level factors determine differences in
levels of tolerance. The most important finding in this chapter is the critical role that education
plays in raising levels of tolerance: higher levels of education lead to higher levels of tolerance.
Is tolerance necessary for democracy? While most citizens support democratic norms
worldwide, this same group of citizens is unlikely to extend democratic rights to disliked
groups (Peffley and Rohrschneider 2003). Consequently, if disliked groups or minorities
generally are denied the right to express themselves freely and their rights are not guaranteed,
how can one guarantee the continuation of democracy over the long term (Seligson and Crdova
1993)?2 For this reason, tolerance is necessary for the persistence of democracy as a system of
government. Moreover, tolerance is necessary for democracy not only because it provides an
environment of political freedom and diversity as well as the opportunity for political
oppositions representation, but also because it may help promote the norms that make for a
peaceful coexistence among divergent groups living together under a single political system.
In this chapter, the main objective is to examine the importance of political tolerance for
democracy. First, the differences in levels of political tolerance across countries will be explored,
focusing on Latin American and Caribbean contexts.3 In order to accomplish that objective, an
index of political tolerance will be developed to measure the acceptance by citizens of the rights
1
This research project benefited from the financial support provided to the AmericasBarometer and a graduate
fellowship for the authors education by USAID. The author is also grateful for the computer-equipped office space
provided by the Center of the Americas and the Provost at Vanderbilt University. The author is highly indebted to
her Latin American Public Opinion Project (LAPOP) colleagues for constantly given advice and guidance
throughout the process of working on this research, and especially to Professor Seligson for his unconditional
support and for making this research possible in the first place.
2
The point of departure for the theoretical basis for a stable democracy is Seligson and Crdovas (Seligson and
Crdova 1993; Seligson 2000) study. Nations can become authoritarian if tolerance is low even if system support is
high, according to that research.
3
North American countries are included in the first part of the analysis to contrast differences in levels of political
tolerance between established democracies and emergent democracies.
195
of citizens who are persistent critics of their system of government. Next, the factors that
influence political tolerance will be analyzed, focusing on the critical role that education plays for
raising levels of political tolerance. The last section of this chapter will be devoted to another
dimension of tolerance, what is termed here as social tolerance, measured by the expressed
level of acceptance of the political rights of homosexuals. The chapter concludes with the
implications of the findings.
Political Tolerance
The literature on political tolerance points to the difficulty posed by its measurement. For
instance, Stouffer (1955) focused on measures related to the willingness of citizens to extend civil
rights to those who have different political views, such as socialists, atheists, and communists. He
predicted that levels of tolerance would increase noticeably in future years. Sullivan et al. (1982)
criticized Stouffers approach by finding that increases in tolerance were merely an illusion
because in reality the dislike toward those specific groups (especially communists) decreased,
while the level of generalized intolerance remained unchanged. Prothro and Grigg (1960) studied
the tolerance of the majority toward minorities and concluded that majorities strongly support the
principles of tolerance but fail to do so with respect to specific minority groups (e.g.,
communists, blacks, gays). Thus, finding the most appropriate way to measure political tolerance
has remained a challenge for social scientists (Dahl 1971; Gibson and Bingham 1982; McClosky
1983; Gibson 1992a). 4
The AmericasBarometer 2006-2007 round of surveys carried out by the Latin American
Public Opinion Project (LAPOP) has its own measures of political tolerance, which have been
found to be reliable in prior studies, and have the advantage as well of providing comparisons
across time and countries. The 2006-2007 survey covers four aspects of democratic rights, and
rather than focusing on a specific minority group (e.g., communists), it looks at the rights of
people who persistently only say bad things about their countrys form of government. The
following questions were asked in all countries; response categories ranged from 1 firmly
disapprove to 10 firmly approve:
D1. There are people who speak negatively of the countrys form of government, not just the incumbent
government but the system of government. How strongly do you approve or disapprove of such peoples right to
vote?
D2. How strongly do you approve or disapprove that such people be allowed to conduct peaceful
demonstrations in order to express their views?
D3. How strongly do you approve or disapprove of such people being permitted to run for public office?
D4. How strongly do you approve or disapprove of such people appearing on television to make speeches?
4
For key studies on political tolerance, see (Gibson 2006) (Gibson 2005a) (Gibson 2005b) (Gibson 2002) (Gibson
1987) (Gibson 1989) (Gibson 1992b)(Gibson and Duch 1991) (Gibson and Duch 1993).
196
The advantage of this methodology is that it does not exclude those who do not select a
least-liked group; neither does it link the answer to a specific group.5 Its shortcoming is that it
focuses only on those who always speak negatively about the form of government. A
consequence of this measure is that the people who disagree with the form of government tend to
be more tolerant toward those who think the same way they do. Yet, that defect is easily resolved
by controlling for the political preferences of the respondents (Seligson, et al. 2006).
The United States 75.7
Canada 73.6
Jamaica 72.7
Venezuela 66.5
Guyana 64.3
Uruguay 62.4
Costa Rica 62.2
Haiti 62.1
Paraguay 57.4
Chile 56.3
Mexico 56.2
El Salvador 55.8
Brazil 55.1
Peru 53.6
Nicaragua 53.5
Guatemala 52.7
Colombia 51.8
Panama 48
Ecuador 46.8
Honduras 46.2
Bolivia 43.9
20 30 40 50 60 70 80
Political Tolerance
Error bars: 95% CI
Figure IX-1. Political Tolerance by country
Figure IX-1 shows that on a scale from 0 to 100 the average level of political tolerance is far
higher in the United States, Canada, and Jamaica than it is in any of the Spanish-speaking or
French-speaking countries in the study. In contrast, Panama, Ecuador, Honduras, and Bolivia
demonstrate the lowest levels of political tolerancebelow 50 points on scale from 0 to 100.
These results may suggest that countries that are stable democracies enjoy higher levels of
political tolerance, possibly because when civil liberties have been in place for longer periods,
citizens have more opportunities to apply democratic norms to disliked opponents (Peffley and
Rohrschneider 2003: 245). For instance, the United States and Canada have a long history of
democratic governance (i.e., from 1787 and 1867 respectively6). However, these results also
suggest that even though established democracies enjoy higher levels of political tolerance, they
5
The current conventional format of tolerance items is to ask respondents to select their least liked group. In a
LAPOP study of Nicaragua, nearly half of the respondents refused to select a group.
6
http://www.infoplease.com/countries.html
197
are far from achieving universal agreement on tolerance. Thus, even among the countries at the
top of the group, a culture of political intolerance is present among an important segment of the
population (Gibson 1992b).
Other cases in the sample that enjoy high levels of political tolerance are Costa Rica and
Venezuela7 which, like the United States and Canada, have the longest democratic tradition of
any of the Latin American and Caribbean cases in the sample (i.e. from 1949 and 1959
respectively). However, the order of the other countries is difficult to explain. Haiti has relatively
high levels of political tolerance, but does not have a British colonial background and no history
of democracy. At the other end, Bolivia is very low, perhaps as a consequence of being the most
deeply ethnically divided of any country in the sample.
Why Does Political Tolerance Differ Across Countries?
It is noteworthy that Jamaica (73) and Guyana (64) compared to other non-North
American countries in the study enjoy high levels of tolerance despite their short history of
democratic performance (i.e., from 1962 and 1992 respectively) and lower levels of economic
and political development similar to countries in Latin America. Why, then, are their levels of
political tolerance relatively high? Latin America and Caribbean countries have been
characterized by a long history of political and economic instability. At the same time,
widespread inequality has been present in the region (Patrice 2003). Many scholars contend that
one explanation for the differences in political and economic situations between developed and
developing countries in the Americas is cultural legacies of colonization. The success of northern
hemisphere countries, namely the United States and Canada, is attributed to its English-Protestant
tradition that generated stronger political and economic institutions (Engerman and Sokoloff
2002). In contrast, the underdeveloped nature of Latin American countries is attributed to the
unequal institutional framework by the Spanish conquest (Yeager 1995).8
Do these cultural and institutional factors relate to the differences observed in political
tolerance? One common characteristic that the U.S., Canada, Jamaica, and Guyana share is their
British colonial heritage. Therefore, it seems that one factor to be related to tolerance may be
colonial heritage (Seligson 1987).
Various authors concur that culture plays an important role in contributing to both
economic development and democracy (Harrison and Huntington 2000). Similarly, Lawrence
Harrison (1985) argues that Latin America is underdeveloped and unstable because its culture
was mainly influenced by the antidemocratic values of traditional Spanish. Harrison also
indicates that these values are transmitted from generation to generation through a socialization
process. Furthermore, Huber et al. (1993) points out that some of the differences between Central
7
Ibid. Because of many political and economic changes that Venezuela has experienced in the past few years, a more
in depth explanation of Venezuelas high political tolerance will be presented at the end of this section.
8
The encomienda system established by the Spaniards in Latin America is believed to have contributed to the
institutionalization of inequality, meaning that all the political and economic power belonged to the elites while the
masses derived little benefit.
198
America and Caribbean countries were caused by their development in the 1930s. According to
these authors:
British colonialism was important because it was an alternative to the Central

American pattern of landlord or military control of the state, and thus an alternative
to the use of labor unions and allied political parties. Consequently, the 30s marked
the beginning of organized political life, and opened the way for the subsequent
consolidation of civil society in the West Indies, whereas in Central America they
solidified the pattern of the primacy of the coercive apparatus of the state and of state
control over the repression of civil society, exercised either by landowner-military
collations or the military alone (79).
Therefore, higher levels of political tolerance in Jamaica and Guyana may be attributed to
the processes of their colonization. Even though these accounts do not provide a complete
explanation, they do offer some insight into the important role that colonialism played in these
countries. Nevertheless, more empirical evidence needs to be obtained to explain the differences
in political tolerance in these countries.
Which Individual-level Factors Influence Differences in Political Tolerance?
The primary research question in this chapter is to determine and explain the factors that
account for the differences in levels of political tolerance. As suggested at the beginning of this
chapter, tolerance matters because it legitimizes the democratic system by guaranteeing the rights
of the minorities. As Gibson (1992b) points out, tolerance matters because it is connected to a
set of beliefs about the legitimacy and appropriateness of self-expression (343). Thus, it is
central to this study to evaluate the factors that contribute to levels of political tolerance.
In order to determine which factors influence political tolerance, the sample of countries
included in this study is divided into three regions: 1) North America and the Caribbean; 2)
Central America; and 3) South America.9 Differences between men and women and their levels
of education will be examined, beginning with the role of education in raising levels of political
tolerance and in greater acceptance of disliked groups. Education is believed to be a critical
element because it enhances political tolerance even among those who have negative feelings
toward a specific group (Lawrence 1976; Bobo and Licari 1989; Golebiowska 1995). Therefore,
it is expected that while higher levels of education should lead toward higher levels of political
tolerance, it is the quality of the education that may well have a greater impact on political
tolerance. Similarly, while various studies show that women tend to have lower levels of political
tolerance than men, there is reason to believe that education will have an impact on womens
tolerance and that differences between both men and women will be mitigated by higher levels of
education. Thus, it is expected that women will show lower levels of political tolerance than men,
yet the higher the levels of education women achieve, the lower their deviation from men.
9
North American countries are included in the first part of the analysis to contrast the differences in levels of
political tolerance between established and emergent democracies.
199
North America and the Caribbean
Figures IX-2 and IX-3 show some variation from the standard pattern of relationship
between education and political tolerance across various North American and Caribbean
countries. The results for Canada and the United States10 show that men are more tolerant than
women at all levels of education. The remaining cases conform to the general pattern, except that
in Haiti, education seems to play no role for women, and secondary education has the strongest
positive effect for men. Guyana is the one case in which gender does not distinguish respondents
by their levels of tolerance. It is worth noting that there are no statistically significant differences
between men and women in Mexico, Haiti, and Guyana whereas the United States and Canada11
show significant differences at the highest level of education.
Canada The United States

85 85
Male Male
Female Female
80
80
Political Tolerance
Political Tolerance
75
75
70
70
65
60 65
55
60
Secondary Tertiary
Secondary Tertiary
Education
Education
Mexico Dominican Republic
65 75
Male
Female
70
60
Political Tolerance
Political Tolerance
65
55
60
50
55
Male
Female
45 50
None Primary Secondary Tertiary None Primary Secondary Tertiary
Education Education
Figure IX-2. Political Tolerance by sex and education: North American and Caribbean countries
10
The primary education category was excluded in the U.S. and Canadian graphs since there were only 2 and 6
cases, respectively. The no education category was excluded in the case of Guyana with only 7 cases. The reason
for this decision is that means are unreliable when sample sizes are very small.
11
The Dominican Republic shows statistically significant differences throughout all levels of education except at the
highest level.
200
Haiti Guyana
80
Male Male
Female Female
64
75
70
Political Tolerance
Political Tolerance
62
65
60
60
55
58 50
45
None Primary Secondary Tertiary Primary Secondary Tertiary
Education Education
Figure IX-3. Political Tolerance by sex and education: Haiti and Guyana
These results do not support the idea that differences between men and women are
mitigated by higher levels of education. Even when women achieve higher levels of education,
they still exhibit lower levels of tolerance compared to men. Some research suggests that women
tend to be more conservative than men. Ewa Golebiowska (1999) points out that women are less
tolerant because of their weaker commitment to abstract norms of democracy, partially due to
their lower levels of political expertise. Women are also less willing to put up with groups
outside the political mainstream because they perceive more threat from such groups, partially
deriving from their greater commitment to moral traditionalism and intolerance of uncertainty
(58).
Central America
Figure IX-4 show a clearer pattern of the relationship between education and political
tolerance in the Central American region in comparison to the region previously examined. Costa
Rica, El Salvador, Nicaragua, and Panama illustrate that political tolerance increases with higher
levels of education. However, this pattern disappears among women in Honduras.
201
Costa Rica El Salvador

80 75
Male Male
Femal Femal
70
Political Tolerance
Political Tolerance
65
60 60
55
50
40 45
Education Education
Nicaragua Panama
Male 57 Male
63
Female Female
54
60
Political Tolerance
Political Tolerance
51
57
48
54
45
51
42
48
Education Education
Guatemala Honduras
56
Male Male
Femal Female
55 55
Political Tolerance
Political Tolerance
54
53 50
52
51 45
50
Education Education
Figure IX-4. Political Tolerance by sex and education: Central American countries
202
Similarly, Panama shows a relationship between education and political tolerance with the
exception of the highest levels of education, at which women have more tolerance than men.
However, the difference is not statistically significant. The only Central American country that
demonstrates marked differences in tolerance between men and women is El Salvador. One could
speculate that while acquiring higher levels of education influences political tolerance, it is the
quality of the education that has a greater impact on political tolerance (Moreno and Seligson
2006). Guatemala is the single country in the region without a clear pattern for women, but in the
case of men it conforms to the general pattern.
South America
Figure IX-6 indicates that political tolerance increases with higher levels of education in
Colombia, Bolivia, Chile, and Paraguay. This relationship appears to hold among men whereas
among women, the relationship disappears with the exception of Chile. In Venezuela, Ecuador,
and Peru (IX-5) education plays no role for men or women. Colombia is the only country in the
region that shows statistically significant differences by sex.
Venezuela
75
Male
Female
70
Political Tolerance
65
60
55
Education
Peru
Ecuador 70
Male
51
Male Female
Female 65
48
Political Tolerance
45
Political Tolerance
60
42
55
39
50
36
33
45
30
Ninguno Primary Secondary Tertiary
None Primary Secondary Tertiary Education

Education
Figure IX-5. Political Tolerance by sex and education: Venezuela, Ecuador, and Peru
203
Uruguay Brazil
Male 65 Male
Female Female
80
Political Tolerance
Political Tolerance
60
70
60 55
50
50
Primary Secondary Tertiary Primary Secondary Tertiary

Education Education
Paraguay Chile
70
Male Male
70 Female Female
65
65
Political Tolerance
Political Tolerance
60
60
55
55
50
50
45
Education Education
Bolivia Colombia
65
Male Male
Female Female
46
60
Political Tolerance
Political Tolerance
44
55
42
50
40 45
40
38

Education
Education
Figure IX-6. Political Tolerance by sex and education: South American countries
204
To sum up, the influence of education on political tolerance is evident among countries
that are considered established democracies. For instance, the United States, Canada, and Costa
Rica show a sharp impact of education and tolerance; the higher the level of education, the higher
the effect on increasing levels of political tolerance. However, there are developing democracies
that also present marked statistical differences between levels of education (i.e., El Salvador,
Nicaragua, Guyana, and the Dominican Republic. We also find this pattern in South America,
except for Paraguay and Peru. These results may suggest that while higher levels of education
should lead to higher levels of political tolerance; it is the quality of education that has a greater
impact on political tolerance (Moreno and Seligson). Furthermore, even with higher levels of
education for females, the differences in political tolerance between men and women are not
eliminated,12 perhaps because, as some research suggests, women are more traditional than men.
Methods and Results
An ordinary least-squares (OLS) linear regression model is used to estimate the effects of
the predictor variables. The model includes a set of theoretically relevant socio-demographic
variables as well as a measure of support for the system.13 The dependent variable, political
tolerance, consists of an index of four questions presented at the beginning of this chapter.
Similarly, a measure of attendance to religious organizations14 is incorporated in the model. It is
expected that regular attendance at meetings of such organizations will have a negative effect on
political tolerance since it relates to support for more traditional values. A measure of
interpersonal trust is also included in the model where those with low levels of interpersonal trust
are expected to have less political tolerance.15 In addition, a measure of the Presidents job
approval is incorporated to control for the opposition to the current government.16 The socio-
demographic variables include measures of age, wealth, and area of residence; a dummy variable
representing females and dummy variables representing primary education, secondary
education, and higher education. This division of levels of education will help predict the
influence that the level of education has on political tolerance. Finally, dummies of all the
countries under examination are also incorporated to render a clear assessment of how political
tolerance differs across countries and across regions.17
The results in Table IX-1 show that education is a significantly positive factor predicting
political tolerance, especially among respondents with secondary and higher education. Similarly,
women tend to have lower levels of tolerance than men. Residing in urban areas as opposed to
rural areas increases levels of political tolerance. For theoretical reasons, system support was
12
In countries that show statistical significant differences between levels of education.
13
The creation of the System Support index is presented elsewhere in this volume.
14
CP6. Meetings of any religious organization, do you attend their meetings at least once a week, once or twice a
month, once or twice a year, or never? The variable was recodified reversing the scale by assigning higher values to
those who attend religious organizations regularly.
15
IT1. Now, speaking of the people from here, would you say that people in this community are generally (1) very
trustworthy, (2) somewhat trustworthy, (3) not very trustworthy or (4) untrustworthy?
16
M1. Speaking in general of the incumbent administration, would you say that the job being done by the President
is: (5) Very good (4) Good (3) Neither good nor bad (fair) (2) Bad (1) Very bad
17
In this section of the chapter, the United States and Canada are excluded from the analysis since the study focuses
on the Latin American and Caribbean contexts.
205
included in the model which shows that system support influences positively political tolerance.
On the other hand, regular attendance to religious organizations, lower levels of interpersonal
trust, and approval of the Presidents job has a negative effect on political tolerance. As
mentioned previously, political tolerance refers to the rights of those who persistently only say
bad things about the countrys form of government. Therefore, the finding that indicates that
those who approve the Presidents job are less tolerant conforms to the general expectation since
they will be less likely to sympathize with the opposition.
Table IX-1. Factors that explain political tolerance: results of the linear regression
B Std. Error Beta t Sig.
System Support .097 .008 .080 11.867 .000
Primary Education .158 .916 .003 .173 .863
Secondary Education 4.040 .934 .072 4.326 .000
Higher Education 9.077 1.009 .128 8.994 .000
Female -2.787 .339 -.050 -8.233 .000
Age -.080 .123 -.004 -.648 .517
Wealth .236 .104 .017 2.268 .023
Urban Areas 1.592 .404 .027 3.938 .000
Attendance to a religious
-.011 .004 -.016 -2.552 .011
organization
Interpersonal Trust -1.062 .189 -.035 -5.621 .000
Presidents job approval -2.756 .198 -.093 -13.916 .000
Mexico -4.243 1.010 -.035 -4.200 .000
Guatemala -6.011 1.073 -.046 -5.602 .000
Salvador -3.379 1.032 -.028 -3.275 .001
Honduras -12.019 1.035 -.098 -11.608 .000
Nicaragua -6.118 1.080 -.048 -5.667 .000
Panama -10.376 1.069 -.084 -9.702 .000
Colombia -7.230 1.047 -.059 -6.908 .000
Ecuador -13.684 1.045 -.111 -13.097 .000
Bolivia -15.500 1.093 -.120 -14.178 .000
Peru -7.407 1.071 -.060 -6.918 .000
Paraguay -1.513 1.062 -.012 -1.424 .154
Chile -4.352 1.023 -.036 -4.254 .000
Uruguay 1.760 1.039 .014 1.695 .090
Venezuela 4.901 1.034 .040 4.741 .000
Dominican Republic .895 1.040 .007 .860 .390
Haiti 5.365 1.174 .039 4.571 .000
Jamaica 12.998 1.089 .096 11.935 .000
Guyana 4.504 1.070 .035 4.208 .000
(Constant) 45.330 1.683 26.930 .000
a Dependent Variable: political tolerance R Square .096
b Country of reference: Costa Rica Adj R Square .094
c Reference level of education: none Estimated std. error 26.697
Ecuador and Bolivia have the lowest levels of political tolerance. Perhaps these findings
could be attributed to the political instability in these countries in recent years. For instance,
206
Ecuador has had seven different presidents in a ten year period. The latest, President Lucio
Gutirrez, was overthrown in April 2005 after the Ecuadorian congress decided to shut down the
Supreme Court.18 In addition, various events marked by corruption (e.g. the Cabrera and
Zambrano cases) had become the trademark of the Ecuadorian political arena. Therefore, it is no
surprise that political tolerance of citizens is low. Nonetheless, a finding that remains puzzling is
the relative high levels of political tolerance that Venezuela reveals despite its drastic political
and economic changes in the past few years.
The case of Venezuela
Venezuela is a federal republic with a multiparty system and one of the most stable
democracies in Latin America since 1959. President Hugo Chvez Frias was elected in
December 1998 and assumed power in early 1999. In July 1999, a constituent assembly was
formed to rewrite the constitution and it was followed by the creation of a constitutional
assembly. Chvez won reelection in December 2006 with 63% of the vote. By 2007, Chvez had
taken important steps to further consolidate his power and create his model of a socialist state. In
January, the National Assembly voted to allow Chvez to rule by decree for a year and half. A
few months later, he shut down one of the main opposition television station, RCTV. In August,
the National Assembly voted to abolish presidential term limits. Chvez has proposed many
reforms to the 1999 Constitution that would further consolidate his power.19
Table IX-2. Factors that explain political tolerance in Venezuela: results of the linear regression
System Support .063 .037 .059 1.690 .091
Primary Education 4.418 4.685 .062 .943 .346
Secondary Education 9.750 4.562 .180 2.137 .033
Female -4.759 1.447 -.088 -3.288 .001
Age .789 .558 .041 1.413 .158
Wealth -.010 .475 -.001 -.020 .984
Urban Areas -1.787 3.652 -.013 -.489 .625
Attendance to a religious organization
.037 .018 .054 2.014 .044
Presidents job approval -3.848 .939 -.160 -4.098 .000
Antichavista sentiment 7.838 2.052 .141 3.819 .000
(Constant) 44.112 7.315 6.030 .000
a Dependent Variable: political tolerance R Square .078
Adj R Square .070
c Reference level of education: none Estimated std. error 26.035
18
Congress decided to close the Supreme Court in December 2004; it was not reinstituted until almost a year later.
19
Tarver and Frederick. 2005. The History of Venezuela. Greenwood Press: Westport, CT.
http://www.venezuelanalysis.com/files/images/2007/10/Constitutional-Reform-Fact-Sheet-2007.pdf
http://www.infoplease.com/ipa/A0108140.html
207
An ordinary least-squares (OLS) linear regression model is used to have a better

assessment of the factors that explain the high levels of political tolerance in this country. The
model includes the same variables as in the previous model, in addition to an index of
antichavista sentiment.20
The results in Table IX-2 show that education is a significantly positive factor predicting
political tolerance, especially among respondents with secondary and higher education. Similarly,
as reported above, women tend to have lower levels of tolerance than men. The strong and
positive significant effect that remains among those with antichavista sentiment even after
controlling for the approval of the Presidents job, indicates that the wording of the tolerance
items is having an impact. That is, the tolerance questions talk about those who only say bad
things about our system of government. Hence, critics of the Chvez style of rule are much
more tolerant of those views than those who are not critical of it.21
Social Tolerance
This section examines another type of tolerance, social tolerance, which is related to the
acceptance of persons and groups perceived as different. In other words, social tolerance involves
full recognition and acceptance of the identity and uniqueness of differences that are seen as not
reducible to invisibility by their bearers (Persell, Green and Gurevich 2001: 208). In this section
of the chapter, social tolerance is referred to as the acceptance of minority groups perceived as
different. Social tolerance will be measured through the level of acceptance of the rights of
homosexuals. Respondents were asked the following question:
D5. And now, changing the topic and thinking of homosexuals, how strongly do you approve or
disapprove of such people being permitted to run for public office?
The contrast between higher levels of political tolerance22 and lower levels of social
tolerance is worthy of note. Figure IX-7 shows that while certain countries demonstrate relatively
high levels of social tolerance (e.g., Panama, Colombia, and Bolivia), they are far from achieving
the same levels as political tolerance, and even then, both types of tolerance remain distant from
being in the positive end of the continuum. Similarly, the figure shows a striking contrast
between North America and Latin America and the Caribbean. Canada and the United States
have, on average, the highest social tolerance levels, with 76 and 70 points respectively on the 0
to 100 scale, while Haiti is the country with the lowest levels of social tolerance.
20
Do you (1) strongly agree, (2) somewhat agree, (3) somewhat disagree, (4) strongly disagree:
C1ch. Do you feel proud of being associated with Hugo Chavez
C3ch. Hugo Chavezs actions make you respect him even more
C4ch. Hugo Chavez weighs the ethical and moral consequences of what he does
C5ch. Hugo Chavez expresses a convincing vision of the future
The index was recodified as a dummy variable: 0. Favor Hugo Chavez 1. Oppose Hugo Chavez
21
For further details on Chvez see: Damarys Canache, From Bullets to Ballots: The Emergence of Popular Support
for Hugo Chvez, Latin American Politics and Society, Vol. 44, No. 1. (Spring, 2002), pp. 69-90.
22
To make comparisons with the Figure illustrating political tolerance across countries, see page 3.
208
Canada 76.2
The United States 69.7
Uruguay 66.5
Brazil 54.1
Mexico 50.4
Chile 48.8
Venezuela 45.6
Costa Rica 41.5
Panama 39.2
Colombia 38.8
Peru 35.0
Bolivia 34.9
Guatemala 34.2
Nicaragua 32.5
Ecuador 32.1
Paraguay 28.6
El Salvador 22.2
Guyana 21.0
Jamaica 19.8
Honduras 18.6
Haiti 9.4
0 10 20 30 40 50 60 70 80
Social Tolerance
Source: AmericasBarometer 2006 Error bars: 95% CI
Figure IX-7. Social Tolerance by country
At the same time, Uruguay and Brazil (i.e., 67, 54 respectively) show high levels of social
tolerance compared to other countries in the region. The majority of countries have levels lower
than 50 points on a scale from 0 to 100, in contrast to only four countries that had levels of
political tolerance below 50 points. The most surprising result is that Caribbean countries that
have the highest levels of political tolerance are those with the lowest levels of social tolerance, at
least when social tolerance is measured in terms of homosexuals. The gap between political and
social tolerance in Caribbean countries may be the result of their historical cultural orientations or
even a reaction to the AIDS epidemic, but at present this is merely speculation.
In this section, the same analysis in the first part of this chapter is done including one
more variable: ethnic identity. This variable is included in the model to determine if ethnic
identity is a significant factor that explains social tolerance. It is expected that Latin American
countries will have lower levels of social tolerance because of the traditional Catholic disapproval
of homosexuality.
209
60
Ethnic Identity 60
White Male
Mestiza Female
Indigenous
50
Black
50
Social Tolerance
Social Tolerance
40
30 40
20
30
10
0 20
Education Education
Figure IX-8. Social Tolerance by education and ethnic Figure IX-9. Social Tolerance by sex and education
identity
Figure IX-8 illustrates that education increases levels of social tolerance.23

Notwithstanding, there are clear differences in levels of social tolerance by ethnic identity.
Whites and mestizos show higher levels of social tolerance, whereas blacks have the lowest
levels. Yet, as will be shown in the regression results, this is largely a result of their average
(low) levels of education and income.
Figure IX-9 shows differences in levels of social tolerance between men and women by
level of education. Here, the results are fascinating since they show that tolerance toward
homosexuals is higher among women than among men whereas in political tolerance, the
opposite is true. One can speculate that women have higher levels of social tolerance probably
because when they were asked if they agree or disagree with homosexuals running for public
office, women might have thought about men in particular. In some ways their answers may
suggest that women are less sensitive about the issue. Still, education has a significant impact on
increasing levels of social tolerance among both men and women. The factors that have an
impact on social tolerance will be analyzed through an ordinary-least squares regression model.
23
These results are based on the pooled sample.
210
Methods and Results
The regression model is the same as the one presented in the previous section of this
chapter, except that it differs by the inclusion of ethnic identity. Four dummies representing
blacks, whites,24 mestizos and indigenous groups are incorporated in the model. The rest of
variables remain the same. 25
Table IX-3. Factors that explain social tolerance: results of the linear regression
System Support .088 .011 .055 8.408 .000
Primary Education -.859 1.183 -.011 -.726 .468
Secondary Education 4.705 1.208 .063 3.896 .000
Female 2.949 .436 .040 6.769 .000
Mestizo 1.320 .571 .017 2.314 .021
Indigenous group 4.576 1.158 .026 3.952 .000
Black -.033 1.064 .000 -.031 .975
Age -.980 .158 -.039 -6.191 .000
Wealth .931 .134 .051 6.942 .000
Urban Areas 2.445 .521 .031 4.689 .000
Attendance to a religious
-.051 .005 -.059 -9.638 .000
organization
Presidents job approval -.781 .255 -.020 -3.066 .002
Guatemala -13.608 1.405 -.078 -9.683 .000
Salvador -26.672 1.311 -.165 -20.350 .000
Honduras -27.180 1.321 -.168 -20.583 .000
Nicaragua -14.152 1.369 -.084 -10.340 .000
Costa Rica -9.764 1.318 -.061 -7.410 .000
Panama -7.154 1.348 -.044 -5.307 .000
Colombia -10.378 1.320 -.063 -7.865 .000
Ecuador -17.503 1.341 -.106 -13.054 .000
Bolivia -13.483 1.433 -.078 -9.412 .000
Peru -16.365 1.354 -.101 -12.089 .000
Paraguay -17.320 1.408 -.104 -12.298 .000
Chile -2.946 1.320 -.018 -2.233 .026
Uruguay 14.893 1.362 .091 10.938 .000
Venezuela -7.839 1.323 -.048 -5.925 .000
Dominican Republic -22.853 1.327 -.139 -17.217 .000
Haiti -33.696 1.765 -.184 -19.096 .000
Jamaica -27.355 1.686 -.151 -16.220 .000
Guyana -26.940 1.448 -.155 -18.599 .000
(Constant) 38.411 2.134 18.001 .000
a Dependent Variable: social tolerance R Square .161
b Country of reference: Mexico Adj. R Square .160
c Reference of level of education: none Estimated std. error 34.14
24
Whites are the reference group.
25
In this section of this chapter, the United States and Canada are also excluded from the analysis since the study
focuses on the Latin American and Caribbean contexts.
211
Contrary to original expectations, the results in Table IX-3 show that even though the effect
on social tolerance of being black is negative, the effect is insignificant when other factors,
especially wealth and education, are included. Therefore, the notion that blacks tend to
particularly dislike homosexuals is not supported by the data. The lower social tolerance found
among blacks was instead a function of their lower levels of education and wealth rather than
their ethnicity. Another unanticipated result is that whites were found to have the highest levels
of social tolerance when the bivariate relationship was presented. Yet, when included in the
regression and controlled by socio-demographic variables and other theoretically related
variables, other ethnic identities express greater social tolerance than whites, especially those
who identify themselves as indigenous. Indigenous groups have less access to education;
therefore, one would have expected that this group would have lower levels of social tolerance.
However, the results show that when controlled for education and overall SES, social intolerance
is lower among the indigenous than among, whites. One could argue that their levels of social
tolerance may reflect greater acceptance of others as a result of having been themselves victims
of intolerance. Nevertheless, further research is necessary to provide more empirical evidence to
explain the reasons indigenous groups have the highest levels of social tolerance.
Other results presented in Table IX-3 indicate that respondents residing in urban areas as
opposed to rural areas are more socially tolerant. Similarly, when education, wealth, and other
variables are controlled for, women tend to show higher levels of social tolerance than men.
System support also influences positively social tolerance. On the other hand, regular attendance
at meetings of religious organizations, lower levels of interpersonal trust and approval of the
Presidents job performance has a negative effect on social tolerance. As mentioned earlier, those
who attend religious organizations regularly tend to be more conservative. Once again, higher
levels of education play an important role. The difference of the effects on social tolerance
between secondary and higher education are sharp. Even though secondary education has a
positive significant effect on tolerance, the effect of higher education is greater by far. These
results reinforce the notion that education is a critical factor in raising levels of social and
political tolerance.
Conclusion
In this chapter, a comparison of political and social tolerance across countries showed that
Jamaica, Guyana, and Haiti compared to other Latin American and Caribbean countries, enjoy
high levels of political tolerance despite their relatively short history of democratic rule and, in
the case of Guyana and Haiti lower levels of economic and political development compared to
many other countries in the study. These results point to the conclusion that perhaps British and
French colonialism contributed to the establishment of certain norms and values that enhance
political tolerance among these countries. On the other hand, these same countries showed the
lowest levels of social tolerance. The study also found that antichavistas in Venezuela are those
with high levels of political tolerance reflecting their greater acceptance of critics of the system.
The most important finding in this chapter, however, is the critical role that education plays in
raising levels of political and social tolerance: higher levels of education lead to higher levels of
tolerance. Therefore, if tolerance is an important goal of democratization, governments ought to
endorse policies that increase the quantity and quality of their educational systems so that
212
tolerance will increase, which in turn will help to ensure the rights of minorities will remain
secure.
213
References
Bobo, Lawrence, and Frederick C. Licari. "Education and Political Tolerance: Testing the Effects
of Cognitive Sophistication and Target Group Affect." The Public Opinion Quarterly 53,
no. 3 (1989): 285-308.
Dahl, Robert A. Polyarchy: Participation and Opposition. New Haven: Yale University Press,
1971.
Engerman, Stanley L., and Kenneth L. Sokoloff. "Factor Endowments, Inequality, and Paths of
Development Among New World Economies." Economia 3, no. 1 (2002): 41-109.
Gibson, James L. "On the Nature of Tolerance: Dichotomous or Continuous?" Political Behavior
27 (2005a): 313-23.
Gibson, James L. "Homosexuals and the Ku Klux Klan: A Contextual Analysis of Political
Tolerance." Western Political Quarterly 40 (1987): 427-48.
. "The Structure of Attitudinal Tolerance in the United States." British Journal of Political
Science 19 (1989): 562-70.
. "Alternative Measures of Political Tolerance: Must Tolerance Be Least-Liked?""
American Journal of Political Science (1992a): 562-71.
. "The Political Consequences of Intolerance: Cultural Conformity and Political
Freedom." American Political Science Review 86, no. 2 (1992b): 338-56.
. "Becoming Tolerant? Short-Term Changes in Russian Political Culture." British Journal
of Political Science 32 (2002): 309-34.
. "Parsimony in the Study of Tolerance and Intolerance." Political Behavior 27 (2005b):
339-45.
. "Enigmas of Intolerance: Fifty Years after Stouffer's Communism, Conformity, and
Civil Liberties." Perspectives on Politics 4 (2006): 21-34.
Gibson, James L., and Richard D. Bingham. "On the Conceptualization and Measurement of
Political Tolerance." The American Political Science Review 76, no. 3 (1982): 603-20.
Gibson, James L., and Raymond M. Duch. "Elitist Theory and Political Tolerance in Western
Europe." Political Behavior 13 (1991): 191-212.
. "Political Intolerance in the USSR: The Distribution and Etiology of Mass Opinion."
Comparative Political Studies 26 (1993): 286-329.
Golebiowska, Ewa. "Individual Value Priorities, Education, and Political Tolerance." Political
Behavior 17, no. 1 (1995): 23-48.
. "Gender Gap in Political Tolerance." Political Behavior 21, no. 1 (1999): 43-66.
Harrison, Lawrence E. Underdevelopment Is a State of Mind: The Latin American Case. Boston:
University Press of America, 1985.
Harrison, Lawrence E., and Samuel P. Huntington, eds. Culture Matters: how values shape
human progress. New York: Basic Books, 2000.
Huber, Evelyne, Dietrich Rueschemeyer, and John D. Stephens. "The Impact of Economic
Development on Democracy." Journal of Economic Perspectives 7, no. 3 (1993): 71-85.
Lawrence, David G. "Procedural Norms and Tolerance: A Reassessment." The American
Political Science Review 70, no. 1 (1976): 80-100.
214
McClosky, Herbert and Alida Brill. Dimensions of Tolerance: What Americans Believe about
Civil Liberties New York: Russell Sage Foundation, 1983.
Moreno, Daniel, and Mitchell Seligson. "Educacin y tolerancia poltica en Bolivia." In La
cultura poltica de los bolivianos. Aproximaciones cuantitativas. Educacin y tolerancia
poltica en Bolivia, edited by Mitchell Seligson and Daniel Moreno. Cochabamba:
Cidadana-LAPOP-USAID, 2006.
Patrice, Franko. The Puzzle of Latin American Economic Development. Edited by 2nd. New
York: Rowman & Littlefield Publishers, Inc, 2003.
Peffley, Mark, and Robert Rohrschneider. "Democratization and Political Tolerance in Seventeen
Countries: A Multi-Level Model of Democratic Learning." Political Research Quarterly
56, no. 3 (2003): 243-57.
Persell, Caroline H., Adam Green, and Liena Gurevich. "Civil Society, Economic Distress, and
Social Tolerance." Sociological Forum 16, no. 2 (2001): 203-30.
Prothro, J.W., and C.W. Grigg. "Fundamental Principles of Democracy: Bases of Agreement and
Disagreement." The Journal of Politics 22 (1960): 276-94.
Seligson, Mitchell A. "Costa Rica and Jamaica." In Competitive Elections in Developing
Countries, edited by Myron Weiner and Ergun Ozbudun. Durham, North Carolina: Duke
University Press 1987.
. "Toward A Model of Democratic Stability: Political Culture in Central America."
Estudios interdisciplinarios de Amrica Latina y el Caribe 11, no. 2 (2000).
Seligson, Mitchell A., and Ricardo Crdova. Perspectivas para una democracia estable en El
Salvador. San Salvador: IDELA, 1993.
Seligson, Mitchell A., Daniel Moreno, Diana Orcs, and Vivian Schwarz-Blum. Auditora de la
democracia: Informe Ecuador 2006. Quito: Cedatos-Gallup International-LAPOP-
USAID, 2006.
Stouffer, Samuel. Communism, Conformity, and Civil Liberties. New York: Doubleday, 1955.
Sullivan, John L., James Piereson, and George E. Marcus. Political Tolerance and American
Democracy. Chicago: University of Chicago Press, 1982.
Yeager, Timothy J. "Encomienda or Slavery? The Spanish Crown's Choice of Labor
Organization in Sixteenth-Century Spanish America." The Journal of Economic History
55, no. 4 (1995): 842-59.
215
Part III. Challenges to
the Rule of Law
X. The Impact of Violent Crime on the Political Culture of

Latin America: The Special Case of Central America
Jos Miguel Cruz
Abstract
The objective of this chapter is to analyze the impact of criminal violence, measured both in
terms of victimization as well as in terms of insecurity, on support for democracy in Latin
America, particularly in Central America, taking into account the degree of consolidation and
stability of democracy in those countries. In other words, the general hypothesis of this work is
that crime erodes democratic political culture in Latin American countries, taking into account,
above all, that many of these countries have a relatively recent history of democratization, as is
the case of most Central American countries: Guatemala, El Salvador, Honduras, Nicaragua
and Panama. The results show that although perceptions of victimization and insecurity play an
important role in eroding democratic political culture in some cases, it is also necessary to
consider citizens perception on government performance regarding public security. This
condition, how the government is perceived, seems to have a more important leverage in some
democratizing countries.
In the early twenty-first century, violence has become a great social problem worldwide.
According to the World Health Organizations World Report on Violence and Health (Krug et.
al. 2002), approximately 1.6 million people lose their lives each year due to violence: violence
is one of the most frequent causes of mortality among people aged 15 to 44. With the exception
of Sub-Saharan Africa, Latin America and the Caribbean stand out as some of the most violent
regions in the world. However, much of the violence in Africa is due to civil wars and ethnic
conflicts, the principle sources of violence in Latin America are social violence and crime.
According to statistics published by the World Health Organization (WHO) on the rate
of homicide in Latin American countries in the 1990s, there were 27 deaths per 100,000
inhabitants, more than twice the world average of 13 deaths per 100,000 inhabitants (Krug et.
al. 2002). In this context, Central America has emerged as the most violent sub-region in Latin
America of late. With homicide rates oscillating between 45 to 55 deaths per 100,000
inhabitants, countries like El Salvador, Guatemala and Honduras are considered countries with
the most severe rates of criminal violence in the world (Gaviria and Pges, 1999).
Very little research has been done on the political impact of common or criminal
violence. From the perspective of political science, violence has not always been examined
under the mode of political violence, that is, as an instrument to accessor vie forpolitical
power, or as a catalyst for the forces that go on to become the State (Bates 2001; Tilly 2003).
Apart from a few recent studies and compilations (see Garland et al. 2003; Karstedt 2006), the
impact of non-political violence on politics has not been explored, particularly in Latin
America. This may be so because historically, when the survival of Latin American
219
democracies has been considered, the economic and social performance of its regimes has been
taken into account more than the capacity to insure public safety.
The surge of criminal violence in societies worldwide has drawn attention to the
possible effects it may have on political processes, and particularly on democracy (Bodemer
2003). Although there are some recent studies on the impact of criminal violence on democracy
and political culture (Prez 2003; Seligson and Azpuru; Cruz 2003; Cruz 2000), most academic
papers focus on the impact of democracy on the functioning of the criminal justice system or on
criminal violence itself (Karstedt and Lafree 2006). Therefore, very little is known about the
impact that criminal and common violence have had on democracy and, above all, on the
processes of democratization that many Latin American countries have been undergoing for
several years.
This becomes more relevant if one considers the fact that violence is an increasingly
pressing problem for many new, stable democracies (LaFree and Tseloni 2006). The objective
of this chapter is to analyze the impact of criminal violence, measured both in terms of
victimization as well as in terms of insecurity, on support for democracy in Latin America,
particularly in Central America, taking into account the degree of consolidation and stability of
democracy in those countries. In other words, the general hypothesis of this work is that crime
erodes democratic political culture in Latin American countries, taking into account, above all,
that many of these countries have a relatively recent history of democratization, as is the case of
most Central American countries: Guatemala, El Salvador, Honduras, Nicaragua and Panama.
In order to perform this analysis, information is drawn from records from the
AmericasBarometer database at Vanderbilt University. This database collects a wide spectrum
of citizen opinions and attitudes from twenty Latin American and Caribbean countries
regarding democracy, politics, corruption, institutional performance, tolerance, criminal
violence, citizen confidence and other issues1.
In addition to theoretical considerations and a review of the conditions of violence in

Latin America, this work presents the results of the study in two major parts. The first focuses
on analyzing the relation between criminal violence and attitudes that support democracy in
Latin American countries as a whole. This is done by means of multivariate analyses of the
results from all countries in the general database of the AmericasBarometer 2006-2007. The
second part specifically focuses on the countries of Central America. The fundamental
objective of this study is to elucidate the impact of criminal violence where crime is particularly
high, and this is the case for the northern Central American region.
Conditions for Democracy, Political Culture, Legitimacy and Violence
Conditions for democracy are those circumstances that make a regime operational and
which prevent it from being overthrown by an authoritarian regime (Rustow 1999; Linz 1989).
1
Countries included in this analysis are: Bolivia, Brazil, Chile, Colombia, Costa Rica, Dominican Republic,
Ecuador, El Salvador, Haiti, Honduras, Guatemala, Guyana, Jamaica, Mexico, Nicaragua, Panama, Paraguay,
Peru, Uruguay, and Venezuela.
220
This study does not analyze the factors that lead to the establishment of democracy, or the
beginning of a democratic transition; rather, it examines the factors that will aid in the
sustainability and the development of democratic regimes once the transition has taken place.
This is the case of Central American regimes with the exception of Costa Rica. Two general
categories explain the circumstances that make a democracy function: first, the focus on
macroeconomic and social situations, which are linked to the economic and political
performance of those regimes; the second, the existence of a group of values, norms and
attitudes held by citizens who support the regimes performance. The latter has to do with
political culture.2
It is not the objective of this study to focus on the impact of a regimes economic
performance on democracy. Studies of economic performance (Lipset 1960/1996; Przeworski
1996; Muller and Seligson 1986) state that when a society produces, for example, economic
growth, reduction of inequality and economic wellbeing, it is less likely that the regime will
rupture, and therefore more likely that its democracy will prosper. Economic performance is,
from this perspective, an important condition for the survival of a regime, particularly if it is
democratic.
This study focuses the role of subjective and socio-cultural factors on the sustainability
and preservation of democracy, in other words, political culture. The concept of political culture
was originally developed by Almond and Verba (1963), who defined it as the set of
psychological orientations towards politics of the members of a society. This was later taken up
by Lipset (1994: 3) who said that democracy requires a culture of support, that a part of the
citizenry and the political elite accept the principle that is the basis for freedom of speech,
communication, association, religion and the rights of association of parties, the rule of law,
human rights, among others.
In its empirical relation to democracy and democratization, however, the notion of

political culture is not equally viewed among researchers given that it implies different values
and attitudes. In this regard, it is possible to identify three principle perspectives on the relation
between political culture and democratization. The first focuses on cultural values and argues
that a stable democracy depends on a satisfactory life, on interpersonal trust, and the
populations resistance to revolutionary change. According to Inglehart (1988), these values can
emerge as the result of economic growth stimulated by capitalism, or as the result of historical
factors related to religion and ideology. The second approach stems from Putnams (1993)
social capital theory, which lists the system of norms of social reciprocity, social trust, and
social participation networks as important factors for the functioning of democracy. This means
that democracy will function and endure in societies in which people trust each other,
participate in organizations and hold values of social reciprocity. The third perspective points to
the link between political culture and democracy, focusing on legitimacy and political support
for the system. This is developed in Lipsets work (1996) on efficiency and legitimacy, and has
been developed empirically by several authors (Diamond 1993; Seligson 2000; Norris 1999).
2
There exists another type of explanation related to political and social structures (see, for example, Lijphart,
quoted in Rustow, 1999), and to historical conditions, as Karl (1990) pointed out. Nevertheless, that type of
explanation is better adapted to transitions than to the processes of democratization. For some excellent studies of
Central American regimes from an historical perspective, see Paige (1996) and Mahoney (2001).
221
According to this perspective, which is the fundamental framework for this study, a
democratic system can only subsist when the populationmasses and elitesare certain that
the existing political institutions are the most adequate for society (Lipset 1960); that is to say,
when most of the population see this regime as the only game in town (Linz and Stepan
1996). This certainty is not created from one day to the next. In democratic regimes, which do
not depend on force, legitimacy is the product of system performance, it emanates from the
perceptions of satisfaction of basic necessities of the majority of the population, and it is related
to the real behaviour of authorities and power groups (Lipset 1960). As Diamond (1993: 13)
points out: legitimacy stems in part from the performance of a democratic regime over time,
but it is also influenced (especially in the early stages of the regime) by the manner in which
specific democratic institutions are articulated with the traditional and legitimate forms of
authority and then by the socialization, the expansion of education, and other types of social
and cultural change.
Diamonds views contain considerations which are relevant in the study of Latin
America, and particularly so in the case of Central America. First, system performance is a
serious challenge to governments that are learning how to function as democratic regimes for
the first time. Given that there are no personal or institutional memories about how democracy
operates, or if there were any, they have vanished for the new generations, the pressure for
results is exceptional, particularly when a link between democracy and satisfaction of needs is
perceived, as in the case of Latin America (Moreno 2001). Second, in societies with a long
authoritarian tradition, such as those in Central America, traditional sources of authoritythe
police and the armed forcesare related to the arbitrary use of force, as are the institutions that
they administer, (Holden 1996). Thus, the assessment of system performance, which is the basis
for legitimacy, is not based solely on economic growth and distribution of wealth as Lipset
originally proposed. It is also constituted by evaluating social reforms by upright and
transparent governments which respect and preserve the rules of the democratic game; and,
within the scope of this study, by their capacity to maintain the law and order (Diamond 1993).
The rampant violence in Latin American, particularly in Central America, can harm the
political culture that is necessary for these democracies to survive. It can affect them in
different ways. First, it can erode the attitudes that lend legitimacy to the system (Perez 2003;
Seligson and Azpuru 2004; Cruz 2003). Violence also erodes social capital, that is,
interpersonal trust, trust in the institutions, and attitudes that favor citizen participation that is
necessary for the construction and maintenance of democracy (Karstedt and LaFee 2006;
Sullivan and Transue 1999). Moreover, violence promotes support for authoritarian responses
which weaken democracy (Kurtenbach 2003). Insecurity and fear of violence which dominates
societies with high levels of crime or in which there is a perception of high crime rates,
produces anxious citizens who, in turn, become bad democrats (Loader 2006, p. 216).
Violence stimulates citizen support for repressive measures which often violate the rule of law,
which is essential for the functioning of a stable democracy.
222
The limited amount of literature about common violence and democracy3 has focused
on the impact of crime on institutional performance and on the establishment of the rule of law
(see Mendez et al. 1999; Aguero and Stark 1998). Beyond measuring the impact on public
legitimacy most studies have analyzed the impact of violence and insecurity on social capital
and tolerance, assuming that they are fundamental values for democratic functioning (Moser
and Holland 1998; Moser and McIlwaine 2001).
It is for these reasons, and because of the lack of empirical studies that portray the link
connecting violence, citizen insecurity and support for the system, that this study intends to test
the theoretical and empirical relation between political culture, support for democracy and
common violence in the countries of Latin America, and particularly in Central American
countries, taking up the analytical framework proposed by Cruz (2000). The premise is that
violence and insecurity erodes support for the system because they reduce credibility in the
systems institutions and they foment attitudes which support alternative authoritarian regimes.
Of course violence and insecurity are not the only factors that generate the lack of
legitimacy and support for the democratic regime. Neither would it be right to say that violence
is necessarily the most important factor in eroding the conditions that support or lead to
democracy. Notwithstanding, it would not be wrong to say that violence and insecurity can play
a fundamental role in destabilizing the democratic processes by debilitating legitimacy, social
capital and the rule of law.
Specifically this study is based on the theoretical framework proposed by Seligson

(2000), according to which high levels of support for the system and a long-term attitude of
political tolerance in the population are fundamental conditions for the maintenance and
survival of a stable democracy. Therefore, the following pages seek to establish the impact of
criminal violence on the combination of these attitudes which sustain a stable democracy.4
Criminal Violence in Latin America
Although the crime rate in Latin America is the highest in the world, violence in Latin
America differs substantially among and within the countries of the region. For years Colombia
has been considered the most violent country in the region with homicide rates of over 80
deaths per 100,000 inhabitants, whereas Chile and Uruguay have rates that are lower than 5
deaths per 100,000 people (De Mesquita Neto 2002). At the end of authoritarian regimes and
3
Most of the literature that links the concepts of violence and democracy focuses on political violence and not on
crime and common violence.
4
This methodology has been chosen instead of simply assessing the impact on support for the political system
because in the latter case, it could be that people support the predominant political system, but at the same time do
not necessarily support a democratic regime. The use of Seligsons model, which combines support for the system
and political tolerance, allows for a better approximation to an unequivocally democratic political culture. On the
other hand, this method also offers an advantage over measuring support for democracy using only one item on the
AmericasBarometer survey (for example ING4: Democracy may have problems but it is better than any other
form of government). The use of this model implies the use of a more complex and standardized measure, and is
less likely to produce the type of slant that appears frequently when using single items.
223
civil wars, Latin American countries have not only discovered new forms of violence, but
their institutions have also begun to develop record systems that have revealed the real
magnitude of the problem of violence.
According to a World Bank study (Moser, Winton and Moser 2004), the homicide rates
in the Latin American region have increased by 50% since the 1980s, and most of the victims of
violence are young people between the ages of 15 and 25. The study points out that the growing
levels of violence have exacerbated social and political problems, and at the same time they
have generated high economic costs in some Latin American countries.
In recent years, some countries have risen to the top of regional violence and crime
statistics. El Salvador, Guatemala and Honduras published information that placed them at the
same level as Colombia and some cities in Brazil. A study funded by the Inter American
Development Bank (IDB) in the mid 1990s showed that El Salvador had homicide rates of over
100 deaths per 100,000 inhabitants (Cruz and Gonzalez 1997) from 1994 to 1997. Another IDB
publication showed that in the years following the Peace Agreements, Guatemala had levels of
almost 150 violent homicides per 100,000 inhabitants (Buvinic et al. 1999). Although some of
these numbers may be overestimated due to the methods of recording (CIEN, 2002; Aguilar
2001), the truth is that in these Central American countries violence is much higher than in the
rest of the region (Arriagada 2001; Wielandt 2005). Table X.1 shows the most recent data on
the homicide rates in Latin America.
Table X.1. Homicide Rates in Latin America and

the Caribbean (the most up-to-date data available)
Country Homicide Rate (per
100,000 inhabitants)
Argentina 4.7*
Bolivia 18.5
Brazil 20.0
Chile 3.0
Colombia 40.0
Costa Rica 6.0
Cuba 6.2*
Ecuador 14.8**
El Salvador 55.5
Guatemala 35.5
Guyana 6.6*
Honduras 41.0
Jamaica 35.0*
Mexico 18.0
Nicaragua 10.0
Panama 11.0
Paraguay 12.6*
Peru 10.3**
Uruguay 4.0
Venezuela 33.0
Sources: Mockus and Acero, (no date);

* Krug et al. (2004);
**Carrin (2004).
224
In Table X.1 it can be seen that the great majority of countries in Latin America have
homicide rates that are equal to or above the global worldwide rate (12/100,000 inhabitants).
Only countries such as Argentina, Chile and Uruguay in the Southern Cone, Costa Rica in
Central America, and Cuba and Guyana in the Caribbean are clearly below this average. As has
already been pointed out, the countries in the northern triangle of Central America are, by far,
the most affected by violence, with homicide rates greater 35 deaths per 100,000 inhabitants;
they are followed by countries that are part of the Caribbean region, such as Mexico, Colombia
and Venezuela, whose rates are also high.
It is worthwhile to point out that the countries in the Andean region also have relatively
high homicide rates given that they have experienced a noticeable increase in recent years.
According to Carrin (2004), the homicide rates in countries such as Ecuador and Peru have
almost quadrupled over a five year span.
The violence that affects many Latin American countries is expressed not only by
homicides but also by the incidence of crime and other expressions of violence. The increase in
juvenile gangs or maras in Central America, which control entire urban areas where the
authorities cannot counter them; the increase of organized crime in countries such as Mexico,
Brazil, Haiti, to levels that openly defy the states; corruption among government officials in
charge of security; and the increase of street crime in some countries that have until now been
considered less violent, such as Argentina or Costa Rica (see for example, Dammert 2000)
generate a feeling of insecurity that affects all citizens.
A survey done by the United Nations Interregional Crime and Justice Research Institute
in the mid 1990s showed that in the six Latin American countries surveyed, over 85% qualified
crimes such as theft, injury and sexual assault as serious or very serious (Gabaldn 2001). In
other words, criminal violence that takes many forms in Latin American countries translates to
citizen insecurity, which in turn affects governability as well as the relations among citizens and
those between citizens and institutions.
Methodological Aspects
This study is based on data generated by the 2006-2007 edition of the Latin American
Public Opinion Projects (LAPOP) AmericasBarometer. It uses the results of a survey of over
30,000 Latin American citizens in twenty countries of Latin America and the Caribbean. The
study uses the results of a survey that is unique in its scope, nature and commitment to the
scientific method5, and begins with the fundamental assumption that peoples beliefs, attitudes
and life-experience are important categories for the analysis of political facts, and that surveys
are a means in which to approach them.
5
Full information regarding LAPOP and the AmericasBarometer can be found at:
http://sitemason.vanderbilt.edu/lapop/about_us.
225
Data
This study uses data from LAPOPs 2006-2007 surveys. The surveys were applied in all
the countries during various periods throughout the year. In each country, national probability
samples were designed in order to precisely represent the population, and a total of 1,500 adult
citizens over the age of 18 were interviewed.6
Items and Variables

Violence, Victimization and Insecurity
In order to assess the impact of violence on the democratic political culture of Central
American citizens, two independent variables will be used. The first variable measures the level
of violence by personal victimization per crime and collects what could be called objective
levels of violence. The second variable measures the level of perceived insecurity as a product
of the violence that prevails in the environment and gathers what might be referred to as the
subjective side of violence.
In the first case, only one item regarding mediation of victimization (VIC1) is used to
determine the levels of crime and violence in the analysis and the models. People were asked
whether they had or not been victims of some violent incident during the twelve months prior to
the interview. In the second case, insecurity was also measured with only one item, AOJ11.
Democratic Political Culture
Generally, the dependent variable is the political culture that supports democracy.
However, given that this culture has many dimensions in which violence can have an impact, a
series of indicators which refer to the factors of a political culture that supports democracy are
used in the general analysis of the data. These dimensions are: a) support for a stable
democracy, following Seligsons model (2000); b) interpersonal trust, measured by way of item
IT1; c) support for electoral democracy as opposed to rejection of authoritarian leaders, through
number AUT1; and d) support for the rule of law, measured by item AOJ8.
There are important theoretical reasons to consider these dimensions part of a political
culture that supports democracy. First, Seligson has presented vast empirical evidence of the
relevance of combining support for the system and interpersonal tolerance as one of the
fundamental actions for democracy. Second, interpersonal trust is reminiscent of relevant social
capital theories (Putnam 1993; Fukuyama 1995), particularly regarding regimes that attempt to
build a social fabric that may promote political participation. On the other hand, rejection of
authoritarian leadership in order to support electoral democracy becomes more important
especially in light of longstanding authoritarian traditions. This dimension is included to
analyze the issue of citizens who support a democratic system in general, trust in others, but are
6
Except for Nicaragua where surveys were carried out starting at age 16.
226
willing to continue supporting authoritarian leadership through elections (Bermeo 2003).

Finally, abiding by the rule of law means that in a democracy all citizens respect the law and
require that institutions do the same (Mendez 2003).
The following table summarizes all the variables used in this study.
Table X.2. Variables in the first part of the analysis

Condition Name of variable Indicator Item on the questionaire
Violence Victimization by crime Percentage of people VIC1: Have you been the victim of
victimized over the a criminal act in the last twelve
last year months?
Insecurity due to crime Insecurity scale AOJ11: Talking about the place or
average (0-100) neighborhood you live in, and
thinking about the possibility of
being the victim of a hold-up or
robbery, do you feel very safe,
somewhat safe, a little unsafe or
very unsafe?
Government Perception that the Government N11: To what extent would you say
Performance government is fighting performance the current administration improves
crime perception average the security of our citizens?
(1-7)
Political Support for stable Combination of high Basic series B1-B6
Culture democracy tolerance and high Basic series D1-D4
support for the system
(0-1)
Interpersonal trust Interpersonal trust- IT1: Now, speaking of people
scale average (0-100) around here, would you say people
in your community are: very
trustworthy, somewhat trustworthy,
a little trustworthy or not at all
trustworthy?
Rejection of authoritarian Authoritarian leader AUT1: There are people who say
leaders rejection average (0- we need a strong leader who does
100) not have to be elected by vote.
Others say that although things are
not working electoral democracy,
that is popular elections, is always
better. What do you think?
Consent to the rule of law Average on the scale AOJ8: In order to capture criminals,
in which the police do you believe the authorities
ignore the law (0-100) always have to respect the law or
can they act outside the law on
occasion?
Empirical Hypothesis
The empirical hypothesis of this chapter holds that victims of crime or violence and
those who show high levels of insecurity as a result of victimization tend to exhibit less support
227
for the political system and less tolerance; in other words, they tend to have a political culture
that does not favor stable democracy. In addition, they more frequently favor the presence of a
strong leader who overlooks electoral democracy and justifies the authorities acting outside the
law as long as they fight crime. These hypotheses will be evaluated using statistical tests of
significance to establish a probable relation between victimization and insecurity and the
aforementioned cultural values.
Victimization, Insecurity and Democratic Political Culture in the Region
The first step is to establish whether criminal violence, measured by the

AmericasBarometer survey as criminal victimization and perceptions of insecurity, affects the
percentage of citizens that have a political culture favorable toward a stable democracy. In
order to do that, a series of regressions were run to establish the impact of insecurity and
violence; control variables such as gender, age, education, family income, and city size (as a
proxy of urbanization) were also added. In addition, a variable tapping government
performance in public security was also included in the models. Given that the impact of
violence and insecurity can be mediated by the functioning of the regime in the public security
sphere, the citizens perception of the performance of the regime was also incorporated into the
models. The assumption behind this is that support for democracy can also be influenced by
citizens perceptions regarding the governments work in the security arena, and cannot only be
influenced by the objective and subjective conditions of security alone.
Hence, attitudes of support for stable democracy, interpersonal trust, support for
electoral democracy (as opposed to approval for authoritarian leadership), and support for rule
of law were regressed on the independent variables depicted above. The results are shown in
Table X.3.
As it can be seen in Table X.3, the variable of crime victimization does not play any
important role on the attitudes leading to supporting a stable democracy (second column).
Crime victimization is not statistically significant and, along with the city size condition, it does
not make a difference in the attitudes of support for stable democracy. We are not surprised by
this finding since crime victims make up only a small part of the population. Conversely,
perceptions of insecurity do have a negative effect upon the attitudes of stable democracy. This
means that in general, people who feel more unsafe because of crime tend to have lower
attitudes of support for stable democracy than the rest of the population. Thus, the problem of
criminal violence seems to have a negative effect on stable democracy only through the
perceptions of insecurity produced by violence, but not through direct victimizationat least
not in the way that it has been measured in the AmericasBarometer survey. However, the
variable that turned out to have a substantive effect and a statistical significant impact on the
support for stable democracy was the assessment of government performance in the public
security area. The latter means that insofar as Latin Americans perceive that their governments
are tackling crime, there is a greater more probability that they will have stronger supporting
attitudes for a stable democracy.
228
Table X.3. Predictors of support for a stable democracy, interpersonal trust, support for electoral
democracy and support for the rule of law in Latin America and the Caribbean, 2006-2007.
Support for a Interpersonal trust Support for Support for the
stable democracy electoral Rule of Law a
democracy
City size .018 1.873** .030* .003

(.010) (.130) (.013) (.009)
Gender (Female) -.074* -1.335 ** .048 .054*

(.030) (.386) (.039) (.027)
Age .006** .145* .011** .008**

(.001) (.013) (.001) (.001)
Family income .047** 1.356** .057** -.027**

(.007) (.085) (.009) (.006)
Education .022** -.116** .042** .014**

(.004) (.050) (.005) (.004)
Crime .000 -.017* .000 -.003**

victimization (.000) (.005) (.001) (.000)
Perception of -.102** -8.837** -.030 -.133**

insecurity (.016) (.209) (.021) (.015)
Assessment on .288** 1.013** .096** .034**

government (.008) (.106) (.010) (.007)
performance in
b
security
Constant -2.209** 61.804** .304* .048**

(.108) (1.370) (.137) (.096)
R2 .090 .116 .023 .018

(Nagelkerke) (Nagelkerke) (Nagelkerke)
X2 o F 1471.25 372.40 287.26 313.91
N 23,923 22,751 22,159 24,167
* p < .05
** p < .001
a
Brazil not included
b
Bolivia not included.
In fact, perceptions of government performance in each of the areas assessed by the

AmericasBarometer survey (tackling poverty-N1, decreasing corruption-N9, democratic
principles promotion-N3, human rights promotion-N10, and tackling unemployment-N12) have
a statistically significant relationship with the support for stable democracy, but in none the
effect is larger with the exception of human rights promotionthan when assessing the
improvement of public security.
229
This is particularly important. It means that having positive perceptions about the
performance of the government in the public security area is more relevant than being a victim
of crime and violence, or than the perceptions of insecurity. When examining the figures by
country, one can find that this relationship is statistically significant in every Latin American
and the Caribbean country included in the sample7. Figure X.1 clearly shows this relationship.
50
Percentage Support toward
40
Stable Democracy
30
20
10
No 1 2 3 4 5 6 7 A Lot
Opinion that Government Improves Public
Security
Source: Americas Barometer 2006 by LAPOP
Figure X.1. Attitudes Supporting Stable Democracy in Latin America and the
Caribbean according to assessment of government performance in public
security, 2006-2007.
All these results, however, should not downplay the importance of perceptions on
insecurity. As it can be seen in the next figure, the percentage of people holding attitudes
supportive of a stable democracy increases insofar as people feel safer from violence. Hence,
people supporting stable democracy goes from 22 percent among those who feel themselves
very unsafe up to 32 percent among those who feel very safe. This relationship is found in the
overall sample, and it is significant in most countries, with the exception of Guatemala, El
Salvador, Ecuador, Bolivia, and Chile.
7
In Bolivia, the question N11 was not included in the questionnaire; hence this country is not included in the
analyses tapping government performance.
230
Support for stable democracy

32%
30%
28%
25%
22%
Very Somewhat Somewhat Very safe

unsafe unsafe safe
Feelings of safety
Figure X.2. Support for a stable democracy in Latin America and the
Caribbean according to citizen perception of safety, 2006-2007
On a related topic, when interpersonal trust is regressed on the violence and insecurity
variables, the results show that all the conditions have a significant effect on interpersonal trust,
even the crime victimization condition. As can be seen in Table X.3, being a victim of crime
and having perceptions of insecurity erode the overall levels of interpersonal trust in the Latin
American region. By the same token, perceptions of government performance in the public
security area also have a positive effect on interpersonal trust. Nevertheless, perception of
insecurity seems to play a more important role in the levels of peoples trust than assessment of
regimes performance in public security.
When examining the effects of these variables using bivariate analyses at the country
level, it can be seen that perceptions of insecurity tend to have a significant effect more
frequently than crime victimization or than perceptions of government performance (see Table
X.4). In all the countries included in the sample, perceptions of insecurity are related to
interpersonal trust with negative signs. Conversely, when exploring the effects of crime
victimization, it turns out that this variable does not have any effect in Bolivia, Chile, Costa
Rica, Jamaica, or Panama; whereas the effect of perceptions of government performance turns
out to be insignificant in Costa Rica, the Dominican Republic, Ecuador, Nicaragua, and
Panama. In short, perceptions of insecurity are the most important eroding condition for
interpersonal trust.
231
Table X.4. Relationships of statistical significance between crime victimization, perceptions of

insecurity and perceptions of government performance in security, and interpersonal trust
according to country, Latin America and the Caribbean 2006-2007 (bivariate analyses)
Independent Variables
Government
Country Victimization Insecurity
performance
Bolivia ** ---
Brazil ** ** *
Chile ** *
Colombia * ** *
Costa Rica **
Dominican Republic * **
Ecuador * **
El Salvador ** ** **
Guatemala ** ** *
Guyana * ** **
Haiti * ** *
Honduras ** ** **
Jamaica ** *
Mexico ** ** **
Nicaragua * **
Panama **
Paraguay * ** *
Peru * ** **
Uruguay ** ** **
Venezuela ** ** **
Dependent Variable: Interpersonal Trust.

* p < .05
** p < .001
When regressing support for electoral democracy (or rejection of authoritarian

leadership) on violence and insecurity variables, the coefficients of the logistic regression
shown in Table X.3 point out that only the perception of regimes performance plays an
important role in supporting electoral democracy and rejecting authoritarian leadership.
Victimization events and the levels of insecurity do not seem to affect support for electoral
democracy, at least in statistical terms.
Although the effect of perception of government performance is significant in the entire

region, in some countries the relationship is non-existent. The way in which citizens perceive
their governments work in the security arena is important in order to reject authoritarian
leaders in Guatemala, Nicaragua, Panama, Colombia, Haiti, the Dominican Republic,
Venezuela, Chile, Paraguay, and Uruguay; but it does not seem to have the same importance in
Mexico, El Salvador, Honduras, Costa Rica, Ecuador, Peru, Brazil, Jamaica, or Guyana.
Therefore, it is sounder to think that the perceptions about the governments functioning are
relevant to rejecting authoritarian leadership only in some regions of Latin America.
Concerning the support for the rule of law, the results of the regression shows that the
variables of violence and insecurity have an effect upon the backing of the rule of law. Support
for the rule of law was measured using question AOJ8, which explores whether people approve
232
or not the police overlooking the law when prosecuting criminals. This variable proved to be
significantly related to crime victimization, to perceptions of insecurity, and to perceptions of
government performance. Results from the logistic regression (Table X.3, column 5) point out
that being a victim of crime in Latin America erodes citizens commitment with the law when
the police fight crime; perceptions of insecurity produce the same effect: to the extent that the
interviewees feel more unsafe, they are more willing to support the police when acting outside
of the law in order to catch criminals. This erodes the rule of law (see Figures X.3 and X.4).
Equally important is the result that reveals that citizens who approve the governments
functioning are more willing to demand that the police follow the laws and rules. In other
words, people who assess negatively the performance of the government are more ready to
consent to police acting outside of the law. Such results have important implications for the
discussion of support for democracy and the rule of law. This means that it is crucial that
governments be perceived as able to deal with security issues so that citizens can commit
themselves to the rule of law.
60
50
Support Rule of Law (Scale 0-100)
40
30
57.7
48.9
20
10
0
No Yes
Crime Victim
Error bars: 95% CI
Figure X.3. Support for the Rule of Law in Latin America and the Caribbean
and crime victimization, 2006-2007.
233
62
Support Rule of Law (Scale 0-100)
60
58
55
52
Very unsafe Somewhat Somewhat Very safe

unsafe safe
Figure X.4. Support for the Rule of Law in Latin America and the Caribbean,
and perceptions of insecurity, 2006-2007.
Although support for the rule of law is influenced by crime victimization, perceptions of
insecurity, and perceptions of government performance in the entire region8, it is necessary to
explore whether these variables have the same effect at the national level. Table X.5 presents an
overview of the bivariate relationships by country. As shown, in Costa Rica, Panama, and the
Dominican Republic, none of the independent variables turn out to be statistically significant.
However, in the other countries, at least one variable came out significant to predict the support
for the rule of law. The perception of insecurity again, seems to play an important role in most
of the countries: citizens tend to reject institutional behavior that is outside of the law insofar as
they feel safe from violence.
8
With the exception of Brazil because the item AOJ8 was not included in their questionnaire.
234
Table X-5. Statistically significant relationships between crime victimization, perceptions of

insecurity and perceptions of government performance in security, and the support for the rule
of law according to country, Latin America and the Caribbean 2006-2007 (bivariate analyses)
Independent variables
Government
Country Victimization Insecurity
Performance
Bolivia * * ---
Chile ** * *
Colombia *
Costa Rica
Dominican Republic
Ecuador * **
El Salvador **
Guatemala ** ** *
Guyana ** ** **
Haiti ** **
Honduras ** * *
Jamaica * **
Mexico ** *
Nicaragua **
Panama
Paraguay * **
Peru *
Uruguay *
Venezuela ** * *
Dependent variable: Support for the Rule of Law.

* p < .05
** p < .001
In sum, crime victimization and perceptions of insecurity affect attitudes of support for
democracy, namely those which maintain a stable democracy, bolster interpersonal trust, reject
authoritarian leadership (and support electoral democracy), and encourage the rule of law. This
is especially true in the case of perception of insecurity concerning crime and violence.
Insecurity affects, in substantive and significant terms, the attitudes that are essential for
maintaining a stable democracy; it also affects interpersonal trust, the beliefs that authorities
must not disobey the law, and the support for the procedures of electoral democracy.
Crime victimization, although important in some cases, does not seem to have the
relevance found in the perceptions of insecurity in predicting democratic political culture
attitudes. Victimization has only a crucial effect when related to the rule of law, and up to some
point when related to interpersonal trust. Nevertheless, victimization does not have significant
effects in rejecting authoritarian leaders, nor in the overall attitudes that are useful for a stable
democracy. Two possible explanations can be drawn from this. First, victimization affects only
a small portion of the population. Even in the most problematic societies victimization does not
reach everyone, so the effect of these kinds of events are limited to a tiny fraction of the
population. Second, victimization has been measured with some limitations because surveys do
not capture all kinds of violence occurring in a society (think, for instance, in homicides).
235
All in all, these findings suggest that there are ways to explore the impact of violence on
the political culture of democracy. These results seem to imply that perceptions of the
performance of the regime are important when considering support for democracy. This will be
more deeply studied in the next section of the present chapter devoted to study the relationship
between violence and attitudes of support for stable democracy in Central America.
Violence and Attitudes of Support for Democracy in Central America
In this section the impact of violence on the Central American sub-region will be
analyzed. There are two reasons to do so. First, because as has been mentioned before, the
northern Central American region, comprised of Guatemala, El Salvador and Honduras is, for
many reasons, the most violent region in the world regarding criminal violence. At the same
time, Nicaragua, Costa Rica and Panama, with homicide rates of less than 15 per 100,000
inhabitants, suggests that it is also one of the least violent regions in Latin America (see Figure
IX.1). This presents a unique opportunity to compare countries with similar cultures in the same
geographical area but strikingly different in their incidence of violence. Second, Central
America countries have very different histories in terms of their democratic consolidation. In
this region comparisons can be drawn between Costa Rica, the most stable country in Latin
America in terms of democracy (Booth 1998), and Guatemala, El Salvador and Nicaragua,
which have undergone transitions to democracy for the first time in their respective histories
(Torres-Rivas 2001) following years of armed conflicts. Interesting comparisons can also be
made regarding Honduras and Panama which have recently transitioned from military
governments to democracies but which have had certain levels of democracy in the past
(Remmer 1985)9.
Therefore, this has to do with the possibility of comparing countries that show two
conditions that are important in understanding the impact violence has on attitudes that are
behind the consolidation of democracy. The position of these countries regarding these
conditions is presented in Table X.6.
9
To state that countries such as Honduras and Panama have had a certain level of democracy in the past is not a
statement free of controversy. For many authors, these countries have not had democratic regimes in the past, but
rather electoral regimes that have constantly been subjected to supervision by the military or oligarchic groups
(Remmer 1985; Mahoney 2001). These groups have allowed certain liberties and processes of political alternating
to take place but have not achieved the level of consolidation and institutionalization reached in Costa Rica. In any
case, during some periods of the twentieth century, these regimes have differentiated themselves from the more
inflexible authoritarian character of their neighbors Guatemala, El Salvador and Nicaragua.
236
Table X.6. Central American countries according to levels of criminal violence and
stage of democracy
Stable Recent Transition from an Transition together with
Democracy authoritarian government: end of military conflict:
Recovered Democracies. Foundational Democracies
Non violent (rate < Costa Rica Panama Nicaragua
15/100,000 hab.)
Violent (rate > Honduras Guatemala
40/100,000) El Salvador
Thus, criminal violence can have an impact on these countries in differentiated manners,
both in terms of the degree of violence that they face, as well as the function of the
democratic/authoritarian traditions prevalent in these regimes. Countries such as Guatemala, El
Salvador and Nicaragua are considered foundational democracies: they have no institutional
memory of the way a democracy functions and their populations have had to learn to live in the
new political culture that requires the abandonment of authoritarianism.10 It is logical then, to
think that the problems posed by high levels of violence would have a greater impact on the
popular support for a stable democracy. On the other hand, in countries such as Honduras and
Panama which have a previous history of quasi-democratic regimes, the challenge lies in re-
establishing the institutions and values to support democracy. This task can be made less
difficult when a certain type of memory exists, even if it is tenuous. The case model for this is
Costa Rica. Compared to its neighbors, Costa Rica has a longstanding democratic tradition,
with institutions and values that have been firmly established for decades. Therefore, the
challenges posed by violence or economic crises should have less of an impact on the political
culture that supports a stable democracy.
To test the impact of the variables of violence on the support for a stable democracy a
statistical analysis was run on data by means of binary logical regressions, taking as the
dependent variable the support for a stable democracy. Unlike the graphs and analysis of the
previous section, this exercise includes demographic variables such as gender, age, education
and the size of the population where the respondent lives. In addition, two variables related to
the topic of violence were introduced into the models, variables which might play an important
role on the impact on political attitudes. One of these variables is the presence of juvenile gangs
in the respondents neighborhood; the other is the evaluation of government performance in the
area of citizen security. The summary of these results is presented in Table X.7. 11
The figure shows the variables that have enough statistical weight to state that they
influence democratic political culture in each per-country regression that are significant in
predicting support for a stable democracy. Consequently, for example, the results show that in
Guatemala, the size of the city where the respondent lives, the perception of gang presence in
the community and the evaluation of the Guatemalan governments performance in the area of
public security are all conditions that are statistically associated to a political culture that
supports democracy.
10
It is important to recall that these transitions implied the cessation of armed conflicts and the retreat of the
military from politics (Crdova 1996).
11
See appendixes for more on the specific results of the regressions.
237
More concretely, the Guatemala results indicate that people living in small cities or rural
areas, those who perceive gang presence in their communities, and those who evaluate the
Guatemalan governments security policies positively tend to show greater demonstrations of a
political culture that supports a stable democracy. In El Salvador, results clearly show that the
respondents genderthat is, being a womanand the perception of the performance of the
government in office in the area of security are the factors influencing support for a stable
democracy. In Honduras, people who are not as afraid of becoming victims of violence and
those who do not perceive the presence of gangs in their communities show greater support for
a stable democracy as well as the positive evaluation of governmental performance in the area
of public security. In Nicaragua, results again point to the importance of government
performance but they also show that variables such as city size, family income and the
perception of security play an important role as well, although a less significant one.12 In Costa
Rica, aside from positive evaluation of government security policies, personal characteristics
appear to play a greater role: older people with a higher educational level and higher income are
more committed to supporting democratic political culture than the rest of the population.
Finally, in Panama only the perception of security and the positive evaluation of governmental
work contribute to political culture that supports democracy.
Table X.7. Predictors of support for stable democracy in each Central American country
El Costa
Variables Guatemala Honduras Nicaragua Panama
Salvador Rica
City size ** *
Respondent gender ***
Respondent age **
Respondent educational **
level
Respondent income level * **
Degree of victimization
Perception of security ** * **
Gang presence * ** ---
Evaluation of government *** *** *** *** *** ***
performance in the area of
public security
* p< .100
** p < .050
*** p< .001
Note: In the case of Panama, the gang presence variable was not included in the regression because it was not asked about in the local
survey.
These results point to the fact that despite the results of previous section when a
multivariate analysis is run on the conditions that contribute to support for a stable democracy,
personal victimization loses its power of prediction in countries where it had previously
12
We opted for showing the variables that have a significance level of p< .1. Although this is particularly high for
statistical significance in the social sciences, it can be useful for registering those variables that have a more diffuse
effect on democratic political culture and that could be overlooked in the regression. All in all, variables with this
type of significance should be considered with much more care and the best is to give due relevance to variables
that show a level of statistical relevance of p< .05.
238
appeared to be of importance, that is to say in Honduras and in Nicaragua. Additionally when

determining the most important factors in the presence of a political culture that supports
democracy from the perspective of the incidence of violence, the most important condition is
not insecurity or victimization; rather, it is the perception of what governments do to combat
these problems. In other words, support for a stable democracy is more the result of perception
of institutional performance than the absence of these problems.
Actually, results indicate that when the variable of perception of security is introduced
into the models the effect of personal victimization on the attitudes of support for a stable
democracy vanishes in Honduras and Nicaragua where there was a simple relation. This would
suggest that in these countries, from the point of view of support for democracy the perception
of security is more important than the fact of having been a direct victim of violence. People
who have been direct victims of crime tend to feel more insecure and this insecurity shared by
many others affects the attitudes of support towards a stable democracy. Although these results
may seem to contradict the idea that direct violence affects democratic attitudes it must be
remembered that the degree of victimization that is accounted for in this study is partial in that
it refers only to victimization due to economic factors and excludes many of the more serious
types of victimization that take place in Latin America. In truth, the results suggest that the
most important variable regarding citizen attitudes is not violence but the insecurity that it
produces. This is not a new consideration as it has already been pointed out in other studies
(Gabaldn 2004) but it demonstrates that when working on the topic of democratic political
culture, perceptions related to violence cannot be overlooked.
In this context, the findings regarding the importance of perception of government

performance in the area of public security turns out to be revealing. In the previous section it
was seen that in the majority of per-country bivariate analyses perception of security played an
important role in determining attitudes of support for democracy. In other words, in all Central
American countries except Guatemala and El Salvador, insecurity due to crime significantly
reduces support for stable democracy. Nevertheless, the results presented in this section show
that for Nicaragua and Costa Rica the perception of governmental efforts becomes more
important than insecurity, such that when introducing this last variable in the regressions the
effect of the sense of security disappears. As seen in Table X.7 (and appearing in the
Appendixes), the variable of governmental performance evaluation is the most important
condition for predicting support for a stable democracy. This suggests that at the time of
examining the impact of criminal violence on political culture it is important to consider not
only the insecurity engendered by crime, but also and above all, to consider the manner in
which citizens perceive governmental actions in the fight against crime. From the citizen
standpoint what government does to combat violence is more important in supporting a
democratic regime than the existence of violence or insecurity in and of themselves. In other
words, the erosion of a political culture that supports democracy fundamentally comes from the
loss of legitimacy, the result of governmental actions, and not so much from the fact of there
being violence and insecurity or not.
There is no doubt that both aspects are related. In many countries criminal violence
increases or decreases as a result of the actions of government. Nonetheless, this is not always
perceived by people when evaluating the state of their societies. When government
239
performance in terms of security is assessed citizens say that their commitment to democracy
depends more on the systems performance in facing problems than the problems themselves.
This was originally proposed by Lipset and restated by Klingemann (1999) when they stated
that the legitimacy of a democratic system is greatly influenced by its own performance.
The preceding is important because at the time of considering the possibilities of

survival of democracy as a product of citizen support it indicates the need to pay attention to
political institutions and not to the problems. Democracy is maintained, strengthened or
weakened not because of the existence of more or less serious problems in society but because
of the actions the institution undertakes in facing these problems and the manner in which these
actions are perceived by the public. From the perspective of political culture the problem of
violence is a problem for democracy because it affects the perception of system performance,
but in so doing, it also affects the conviction that the rule of law has to be respected, and
weakens the opposition to authoritarian leadership. Therefore, violence is a problem for
maintaining democratic stability.
This can be seen much more clearly in the results of a series of questions included in the
AmericasBarometer with the objective of identifying the conditions in which citizens would be
willing to support a coup detat. Figure X.5 shows these results for the countries of Central
America grouped by the function of level of violence.
Among the conditions listed unemployment, social protest, criminal violence, high
inflation and corruption crime stands out as the condition by which most citizens would
support a coup detat. This is as true for Central American countries that have serious violence
problems as it is for the other countries included in the AmericasBarometer. However, it is in
Guatemala, El Salvador and Honduras where the level of citizen justification for a coup detat
is the highest. Over half of Guatemalans, Salvadorans and Hondurans said they would approve
of a coup detat as a response to the levels of violence that exists in those countries. No other
condition reaches this level of approval and nothing else stands out as notably among the
violent countries and the non-violent ones. In other words, support for the break with
democracy due to criminal violence is the highest in precisely those countries where violence is
a serious problem. In no other circumstance or country is there so much support for a break
with democracy as in those countries where violence prevails.
Under no circumstances should this data be taken as a predictor of a coup detat.

Instead, this data needs to be taken as an indicator of the level of erosion of legitimacy that
regimes face when they co-exist with high levels of criminality. The results shown in the graph
only exemplify what has already been said before: violence affects the perception of the
systems performance and it is this very perception that weakens the prospects of a political
culture that would support democracy. The fact that people might justify a coup detat due to
the high levels of violence fundamentally constitutes a response to modify a regime perceived
incapable of providing security.
240
60%
Gua, ElS, Hon
Rest LAC
50%
40%
Percentages
30%
53.5
47.6
44.5 43.6
20%
29.5
27.6 25.8
24.1
10% 20.7 20.1
0%
Unemployment Social Inflation Corruption Criminal
Protests Violence
Conditions for justifying a military coup

Figure IX.5. Conditions that might justify a coup detat in Central American countries with
problems of violence compared to the rest of countries, 2006-2007.
Conclusions
Criminal violence affects the prospects of a stable democracy. This is particularly true in
most of the countries in Central America. Violence creates insecurity and contributes to
convincing citizens that the institutions in office do not provide adequate security nor do they
do enough to combat crime. This should not be a problem of legitimacy because in a democracy
institutions can be modified, reformed or strengthened, and their representatives can be called
to account, relieved or backed following a series of democratic procedures. What is certain is
that data indicate that when citizens live in an environment of insecurity, and when they do not
have a positive evaluation of the performance of the institutions that are in charge of security,
an erosion in the attitudes that support a stable democracy becomes apparent to the degree that
in countries where violence is one of the most serious social problems, over half of the people
justify a coup detat as a manner of dealing with the problem of criminal violence.
241
The impact of violence on democratic political culture, therefore, is not related only to
the prevalence of crime and insecurity but above all to the conditions that erode the legitimacy
of the institutions that are in charge of providing security and as a result, cause people to detach
themselves from democratic values, ignore the rule of law and support authoritarian
alternatives. This is not to say that the most violent countries are irredeemably on the way to the
collapse of their proto-democratic regimes. It does mean that criminal violence, particularly of
the most severe kind that translates into generalized insecurity, constitutes a clear threat to the
processes of democratic consolidation.
This is particularly true in the countries that are most affected by violence in Central
America. Having examined the transitional character of most of these countries, there remains
little doubt regarding the impact of violence and insecurity on citizen support for democratic
regimes in those countries. Violence is pernicious to the frail path to democratization in the
Central American countries where democratization is recent, not only because it directly erodes
the rule of law, or because it threatens the citizens most basic human rights but also because it
can destroy the infrastructure of legitimacy that is necessary for new regimes. As stated by
Seligson (2002), Latin America has a long history of instability and authoritarian regressions.
Central America is no exception to this negative tendency; countries in this region were
unstable long before having known democratic regimes. Judging from history in other latitudes
there is no reason to think that the current and hybrid democratic regimes are free from these
setbacks, particularly when violence continues to affect Central American citizens.
242
References
Agero, Felipe and Stark, Jeffrey. (eds). Fault Lines in Post-Transition Latin America. Miami:
North-South Center Press, 1998.
Aguilar, Jeannette. Metodologas para la cuantificacin del delito. San Salvador: Fundacin de
Estudios para la Aplicacin del Derecho (FESPAD), Centro de Estudios Penales de El
Salvador (CEPES) (2002).
Almond, Gabriel and Verba, Sydney. The Civic Culture: Political Attitudes and Democracy in
Bates, Robert. Prosperity and Violence. The political economy of development. New York:
Norton, 2001.
Bermeo, Nancy. Ordinary People in Extraordinary Times. The Citizenry and the Breakdown of
Democracy. Princeton: Princeton University Press, 2003.
Bodemer, Klaus. El nuevo escenario de (in)seguridad en Amrica Latina Amenaza para la
democracia? Caracas: Nueva Sociedad, 2003.
Booth, J. Costa Rica: Quest for Democracy. Boulder: Westview Press, 1998.
Buvinic, Mayra; Morrison, Andrew; and Shifter, Michael. La violencia en Amrica Latina y el
Caribe. In: Fernando Carrin (ed.). Seguridad ciudadana, espejismo o realidad? Quito:
FLACSO Ecuador- OPS/OMS, 2002.
Carrin, Fernando. Seguridad ciudadana en la comunidad andina. Iconos 18 (2004): 119-
129.
Centro de Investigacion Econmicas Nacionales. (CIEN). Estudio sobre la magnitud y el costo
de la violencia en Guatemala. Guatemala: CIEN, 2002.
Crdova Macas, Ricardo. Executive-Legislative Relations and the Institutionalisation of
Democracy. In: Rachel Sieder (ed.). Central America: Fragile Transition. London:
Macmillan Press, 1996.
Cruz, Jos Miguel. Violencia y democratizacin en Centroamrica: el impacto del crimen en la
legitimidad de los regmenes de posguerra. Amrica Latina Hoy 35 (2003): 19-59.
. Violencia, democracia y cultura poltica. Nueva Sociedad 167 (2000): 132-146.
Cruz, Jos Miguel y Gonzlez, Luis Armando Magnitud de la violencia en El Salvador.
Estudios Centroamericanos 588 (1997): 953-966.
Dammert, Luca. Violencia criminal y seguridad pblica en Amrica Latina: la situacin en
Argentina. Serie polticas sociales No. 43. Santiago de Chile: CEPAL, 2000.
De Mesquita Neto, Paulo. Crime, Violence, and Democracy in Latin America. Paper
presented in the Conference Integration in the Americas. Albuquerque, New Mexico:
April 5, 2002.
Diamond, Larry. Introduction: Political Culture and Democracy. In: Larry Diamond (ed.).
Political culture and democracy in Developing Countries. London: Lynne Rienner
Publishers, 1993.
Fajnzylber, Pablo; Lederman, Daniel and Loayza, Norman. Crimen y violencia en Amrica
Latina. Bogot: Alfa y Omega, 2001.
Fukuyama, Francis. Trust. the social virtues and the creation of prosperity. New York: Free
Press, 1995.
243
Gabaldn, Luis Gerardo. Seguridad ciudadana y control del delito en Amrica Latina.
Anlisis y propuestas. El observatorio de Nueva Sociedad. [Puede encontrarse en:
http://www.seguridadregional-fes.org/upload/1772-001_g.pdf] (ltima fecha de acceso: 4
de marzo de 2007).
Gabaldn, Luis Gerardo. Desarrollo de la criminalidad violenta en Amrica Latina: un
panorama. En: Bodemer, Klaus et al. Violencia y regulacin de conflictos en Amrica
Latina. Caracas: Nueva Sociedad, 2001.
Garland, Allison; Holding, Heather A.; Ruthenburg, Meg; and Tulchin, Joseph. (eds.). Crime
and the threat to democratic governance. Washington, D.C.: Woodrow Wilson Center
for International Scholars, 2003.
Gaviria, Alejandro & Pags, Carmen. Patterns of Crime Victimization in Latin America.
Working Paper No. 408. Washington, D.C: Interamerican Development Bank, 1999.
Holden, Robert. Constructing the Limits of State Violence in Central America: Towards a
New Research Agenda. Journal of Latin American Studies 28, 2 (1996): 435-459.
Inglehart, Ronald. The Renaissance of Political Culture. American Political Science Review
8 (1988): 1203- 1230.
Karstedt, Susanne. Democracy, values, and violence: paradoxes, tensions, and comparative
advantages of liberal inclusion. The Annals of the American Academy of Political and
Social Science 605 (6) (2006): 50-81.
Karstedt, Susanne y Lafree, Gary. Democracy, Crime and Justice. The Annals of the
American Academy of Political and Social Science 605 (6) (2006): 6-23.
Klingemann, Hans-Dieter. Mapping political support in the 1990s: A global Analysis.. En:
Pippa Norris (ed.). Critical citizens. Global support for democratic governance. Oxford:
Oxford University Press, 1999.
Krug, Etienne H.; Dahlberg, Linda L.; Mercy, James A.; Zwi, Anthony B.; and Lozano Rafael.
(eds). World Report on Violence and Health. Geneva: World Health Organization
Publications, 2002.
Kurtenbach, Sabine. El nuevo escenario de (in)seguridad en Amrica Latina: amenaza para la
democracia?. En: Klaus Bodemer. El nuevo escenario de (in)seguridad en Amrica
Latina Amenaza para la democracia? Caracas: Nueva Sociedad, 2003.
LaFree, Gary y Tseloni, Andromachi. Democracy and crime: a multilevel analysis of
homicide trends in forty-four countries, 1950-2000. The Annals of the American
Academy of Political and Social Science 605 (6) (2006): 26-49.
Linz, Juan and Stepan, Alfred. Hacia la consolidacin democrtica. La Poltica 2 (1996): 29
49.
Linz, Juan. La quiebra de las democracias. Madrid: Alianza Editorial, 1989.
Lipset, Seymour M. The Social Requisites of Democracy Revisited. American Sociological
Review 59, 1 (1994): 1-22.
. Political man. London: Mercury Books, 1960.
Loader, Ian. Policing, Recognition, and Belonging. The Annals of the American Academy of
Political and Social Science 605 (6) (2006): 210-223.
Londoo, Juan L., Gaviria, Alejandro & Guerrero, Rodrigo (eds.). Violencia en Amrica
Latina. Asalto al desarrollo. Washington, D.C.: Banco Interamericano de Desarrollo,
2000.
244
Mahoney, James. Radical, reformist, and aborted liberalism: Origins of national regimes in
Central America. Journal of Latin American Studies 33, 2 (2001): 221-256.
Mndez, Juan. Problems of Lawless Violence: Introduction. In: Juan Mndez, Guillermo
ODonnell and Paulo Sergio Pinheiro (ed.). The (Un)Rule of Law & the Underprivileged
in Latin America. Notre Dame: University of Notre Dame Press, 1999.
Mockus, Antanas y Acero, Hugo. Criminalidad y violencia en Amrica Latina. La experiencia
exitosa de Bogot. Seguridad Sostenible 22 (sin fecha).
Moreno, Alejandro. Democracy and Mass Beliefs Systems in Latin America. In: Roderic Ai
Camp (ed.). Citizens views of Democracy in Latin America. Pittsburgh: University of
Moser, Caroline y Holland, Jeremy. La pobreza urbana y la violencia en Jamaica. Washington
D.C.: El Banco Mundial, 1997.
Moser, Caroline y McIlwaine, Cathy. Violence in a Post-Conflict Context: Urban Poor
Perceptions from Guatemala. Washington, D.C.: World Bank Publications, 2001.
Moser, Caroline; Winton, Ailsa y Moser, Annalise. Violence, Fear, and Insecurity among the
Urban Poor in Latin America. En: Marianne Fay (ed.). The urban poor in Latin
America. Washington, D.C.: The World Bank, 2005.
Muller, Edward y Seligson, Mitchell A. Inequality and Insurgency. American Political
Science Review 81, 2 (1987): 425-452.
Norris, Pippa. Introduction: The Growth of Critical Citizens? In: Pippa Norris (ed.). Critical
Citizens. Global Support for Democratic Government. Oxford: Oxford University Press,
1999.
Prez, Orlando. Democratic Legitimacy and Public Insecurity: Crime and Democracy in El
Salvador and Guatemala. Political Science Quaterly, 118 (4) (2003). Winter 2003-2004.
Programa de las Naciones Unidas para el Desarrollo. (PNUD). La democracia en Amrica
Latina: Hacia una democracia de ciudadanas y ciudadanos. Buenos Aires: Alfaguara,
2004.
Przeworski, Adam; Alvarez, Michael; Cheibub, Jos Antonio; and Limongi, Fernando. What
makes democracies endure? Journal of Democracy 7, 1 (1996): 39-55.
Putnam, Robert A. Making Democracy Work. Princeton: Princeton University Press, 1993.
Remmer, Karen. Redemocratization and the impact of authoritarian rule in Latin America.
Comparative Politics 17, 3 (1985): 253-275.
Rustow, Dankwart. Transitions to democracy: Toward a Dynamic Model. In: Lisa Anderson
(comp.). Transitions to Democracy. New York: Columbia University Press, 1999.
Seligson, Mitchell A. The Impact of Corruption on Regime Legitimacy: A Comparative Study
of Four Latin American Countries. The Journal of Politics 64, 2 (2002): 408-433.
. Toward a Model of Democratic Stability: Political Culture in Central America.
Estudios Interdisplinarios de Amrica Latina y el Caribe 11, 2 (2000): 1 22.
Seligson, Mitchell y Azpuru, Dinorah. Las dimensiones y el impacto poltico de la
delincuencia en la poblacin guatemalteca. En Luis Rosero Bixby (ed)., Poblacin del
istmo 2000: Familia, migracin, violencia y medio ambiente. San Jos: Centro
Centroamericano de Poblacin, 2000.
Tilly, Charles. The Politics of Violence. Cambridge: Cambridge University Press, 2003.
Torres-Rivas, Edelberto. Foundations: Central America. En: Manuel Antonio Garretn and
Edward Newman, (eds). Democracy in Latin America. (Re)Constructing Political
Society. New York: United Nations University Press, 2001.
245
Wielandt, Gustavo. Hacia la construccin de lecciones del posconflicto en Amrica Latina y

el Caribe. Una mirada a la violencia juvenile en Centroamrica. Serie Polticas Sociales
115. Santiago de Chile: CEPAL, 2005.
246
Appendixes: Binary Logistic Regression Tables According to Country
Table X.A.1: Guatemala. Variables in the Equation(b)

Step size .095 .046 4.303 1 .038 1.100
1(a) q1
.009 .135 .005 1 .946 1.009
q2 .006 .005 1.559 1 .212 1.006
ed .025 .019 1.781 1 .182 1.026
q10 .002 .042 .001 1 .970 1.002
vic2r .000 .003 .028 1 .866 1.000
insecure .004 .003 2.059 1 .151 1.004
aoj17 -.137 .071 3.656 1 .056 .872
n10 .205 .037 30.226 1 .000 1.228
Constant -2.351 .463 25.833 1 .000 .095
a Variable(s) entered on step 1: size, q1, q2, ed, q10, vic2r, insecure, aoj17, n10.
b country country = 2 Guatemala
Table X.A.2: El Salvador. Variables in the Equation (b)

Step size .014 .045 .102 1 .750 1.014
1(a) q1 -.518 .125 17.176 1 .000 .596
q2 -.002 .004 .236 1 .627 .998
ed .012 .017 .463 1 .496 1.012
q10 .011 .034 .097 1 .756 1.011
vic2r .000 .002 .024 1 .877 1.000
insecure .002 .002 1.490 1 .222 1.002
aoj17 -.028 .064 .190 1 .663 .973
n10 .288 .035 66.741 1 .000 1.334
Constant -1.269 .464 7.489 1 .006 .281
b country Country = 3 El Salvador
247
Table X.A.3: Honduras. Variables in the Equation(b)

Step size -.070 .056 1.563 1 .211 .933
1(a) q1 -.087 .136 .412 1 .521 .916
q2 .004 .005 .571 1 .450 1.004
ed .021 .020 1.063 1 .303 1.021
q10 .053 .041 1.707 1 .191 1.055
vic2r -.002 .003 .763 1 .382 .998
insecure .008 .002 10.824 1 .001 1.008
aoj17 .226 .092 6.026 1 .014 1.253
n10
.321 .044 54.327 1 .000 1.379
Constant -3.658 .554 43.552 1 .000 .026
b country Country = 4 Honduras
Table X.A.4: Nicaragua. Variables in the Equation(b)

Step size -.089 .051 3.029 1 .082 .915
1(a) q1 -.027 .136 .040 1 .841 .973
q2 -.001 .005 .044 1 .834 .999
ed -.004 .020 .031 1 .860 .996
q10 .063 .037 2.867 1 .090 1.065
vic2r -.003 .003 1.635 1 .201 .997
insecure .004 .002 2.729 1 .099 1.004
aoj17 .025 .072 .121 1 .728 1.025
n10 .258 .037 48.926 1 .000 1.295
Constant -2.076 .481 18.653 1 .000 .125
b country Country = 5 Nicaragua.
248
Table X.A.5: Costa Rica. Variables in the Equation(b)

Step size .043 .037 1.304 1 .254 1.044
1(a) q1 -.093 .115 .649 1 .421 .911
q2 .008 .004 4.926 1 .026 1.008
ed .034 .016 4.591 1 .032 1.035
q10 .054 .027 3.954 1 .047 1.056
vic2r .001 .002 .115 1 .735 1.001
insecure .002 .002 1.960 1 .162 1.002
aoj17 .059 .055 1.165 1 .280 1.061
n10 .199 .033 36.367 1 .000 1.220
Constant -2.079 .406 26.163 1 .000 .125
b country Country = 6 Costa Rica.
Table X.A.6: Panama. Variables in the Equation(b)

Step size -.064 .044 2.062 1 .151 .938
1(a) q1 -.023 .135 .029 1 .864 .977
q2 .005 .005 1.035 1 .309 1.005
ed .002 .019 .016 1 .899 1.002
q10 .069 .056 1.515 1 .218 1.072
vic2r .000 .004 .000 1 .988 1.000
insecure .006 .002 5.435 1 .020 1.006
n10 .544 .051 114.054 1 .000 1.723
Constant -3.844 .473 66.119 1 .000 .021
a Variable(s) entered on step 1: size, q1, q2, ed, q10, vic2r, insecure, n10.
b country Country= 7 Panama.
249
XI. Corruption and its Impact on Latin American

Democratic Stability*
Dominique Zphyr**
Abstract
The objective of this chapter is to analyze the impact of corruption on citizen attitudes towards
democracy. To this end, the corruption measure developed by LAPOP is used in this study,
which, as will be shown, overcomes many of the weaknesses of the indicators of perception of
corruption. Additionally, there is an assessment of the impact that corruption victimization has
on support for the political system and support for democracy. This is measured by the level of
citizen satisfaction with the way in which democracy functions in practice and support for
democracy as an ideal form of government. The inclusion of most of the Latin American
countries in the AmericasBarometer allows for robust conclusions regarding the impact of
corruption in the region.
Is corruption a challenge for democratization in Latin America and the Caribbean?

Traditionally, studies on corruption have focused on assessing its impact on macroeconomic
development and socioeconomic welfare of developing nations. It has been argued that
corruption affects trade negatively, diminishes incentives for investment, and slows the
development and economic growth (World Bank 1997). However, the impact of corruption on
political development has only recently emerged as a topic for debate in the social sciences. For
some authors, corruption can contribute to political stability (Huntington 1968; Waterbury 1976;
Becquart-Leclercq 1989), given that it can may make state bureaucracies function more
effectively. For example, Becquart-Leclerq (1989) argues that corruption functions like grease
in the gears; it has an important redistributive effect, it is a functional substitute for direct
participation in power, it constitutes the cement between elites and parties, and it affects the
effectiveness with which power is exercised (192). Other scholars have pointed out that
corruption has a negative effect on citizen attitudes towards the legitimacy of the political system,
and it ultimately represents a threat to the consolidation of democratic regimes (Shin 1999;
Camp, et al. 2000; Della Porta 2000; Pharr 2000).1
The study of the consequences of corruption on democratic stability in the Latin

American context is particularly relevant given that global indicators on corruption show that
most of the countries in the region reveal evidence of high levels of corruption. Former
Organization of American States (OAS) Secretary General Cesar Gaviria emphasizes the harmful
effects of corruption and the importance of this issue in the region, pointing out that corruption
is a terrible cancer that threatens the legitimacy of the institutions and the rule of law, and in this
*The author thanks Mitchell Seligson for his suggestions and comments. The errors and omissions are solely the
authors.
** Research Coordinator for the Latin American Public Opinion Project.
1
For a more exhaustive review of literature, see (Seligson 2002; Seligson 2006)
251
sense there is still a long way to go in the Americas.2 Recent data gathering by means of surveys
on political culture which have included items on citizen perception of and/or experience with
corruption have made it possible to provide empirical evidence of the existence of a negative
relationship between corruption and the legitimacy of the political system. Using data from the
World Values Survey (1995-1997) together with those from Transparency International (1997),
Canache and Allison (2005) found that citizen perception of high levels of corruption entails a
lower level of support for the incumbent government and the political institutions in six Latin
American countries. However, they found no evidence that the perception of corruption has any
impact on citizen support for democracy as the best form of government.
Additionally, Seligson (2002, 2006) has developed an innovative methodology to measure

corruption using the surveys of the Latin American Public Opinion Project (LAPOP) as a
database. This new methodology measures levels of corruption based on day-to-day citizen
experiences with bribery instead of perceptions of corruption. Seligson shows that, at the
individual level, experience with corruption translates to less support for the political system in
the six Latin American countries analyzed: Bolivia, Ecuador, El Salvador, Honduras, Nicaragua
and Paraguay.
The objective of the present study is to analyze the impact of corruption on citizen
attitudes towards democracy at a broader level. To this end, the corruption measure developed by
LAPOP is used in this study, which, as will be shown, overcomes the weaknesses of the
indicators of perception of corruption. Additionally, there is an assessment of the impact that
corruption victimization has on support for the political system and support for democracy. This
is measured by the level of citizen satisfaction with the way in which democracy functions in
practice and support for democracy as an ideal form of government. It must be said that the
inclusion of most of the Latin American countries in the AmericasBarometer allows for more
robust conclusions regarding the impact of corruption in the region.
The empirical analysis in this work is based on the 2006-07 AmericasBarometer surveys.
These were carried out person-to-person in twenty-two countries throughout Latin America and
the Caribbean. For these surveys, a stratified, multi-staged, cluster sample was used.
Furthermore, for the purposes of comparison, data from phone surveys in Canada and the United
States gathered by LAPOP were also taken into account. In total, the data base comprises over
30,000 observations.
The principal hypothesis in this chapter is that experience with corruption negatively
affects citizen opinions of democracy and their support for the political system. In other words,
taking into account the idea that the existence of a democratic political culture entails stronger
and more long lasting democracies, this part of the study argues that corruption poses a challenge
for democracy in that it causes citizens to have a less positive view of the performance of
democratic regimes. In the first section of this chapter an appraisal is made of the levels of
corruption in the countries of Latin America and the Caribbean contrasting Transparency
2
Message of the OAS General Secretary, Cesar Gaviria, to a conference on Adjustment of Dominican Penal
Legislation to the Convention Against Corruption Santo Domingo, Dominican Republic, August 30, 2002
252
Internationals Corruption Perception Index with LAPOPs Corruption Victimization Index. The
second section presents a study of the most common forms of victimization by corruption in
Latin America and the Caribbean, that is to say, the places in which citizens are most likely to
become victims of corruption. The third section offers a profile of persons who are the most
likely to become victims of corruption, taking into account individual characteristics. In the final
section the chapters main hypothesis is tested, followed by conclusions.
The Level of Corruption in the Region
How much corruption exists in Latin America and the Caribbean? In order to take a first
look at corruption in the region, Transparency Internationals Corruption Perception Index (CPI),
developed in 1995 is used. The CPI is the oldest and best-known large scale measure of
corruption.3 It is an index scaled from one to ten, in which the value of one indicates a maximum
level of corruption and ten a minimal level of corruption. The 2006 CPI shows that the corruption
average for Latin America and the Caribbean was 3.5. The region presented, on average, higher
levels of corruption than the European Union and Asia, but lower than those for Africa and
Eastern Europe. Nevertheless, according to this index, the Latin American region is home some
of the most corrupt countries in the world, among them Haiti, Venezuela and Ecuador.
Furthermore, twenty-five of the thirty countries in the region for which the CPI was calculated in
2006, show levels of corruption lower than 5, with a third of the countries scores below 3. These
low scores are troublesome when taking into account that Transparency International considers
corruption a serious problem in countries with scores lower than 5, and a rampant problem in
countries with scores of under 3 ( 2006b).
Table XI-1 shows the levels of perception of corruption according to the CPI for the
countries of the Americas that were included in the 2006-07 AmericasBarometer. As can be seen,
there are substantial differences in the perceived levels of corruption among the countries
analyzed here. Chile is the Latin American country with the lowest level of perception of
corruption with a score of 7.3, which places it second, alongside the United States, among the
countries with the least perceived corruption in the Americas. Canada is the country with the
lowest perceived level of corruption in the Americas according to the CPI. On the other hand,
Haiti received the lowest score in the region, (1.8), which places it not only as the most corrupt
country in the Americas, but amazingly, the most corrupt country in the world. According to
Transparency Internationals data, Haiti is in last position among the 163 countries where data
regarding the perception of corruption was collected. Nevertheless, the levels of corruption are
alarming not only in Haiti, but also in another nine of the countries analyzed by the
AmericasBarometer, which show evidence of generalized levels of perception of corruption.
3
The other global index on perception of corruption, which competes with the CPI in popularity, is the Corruption
Control Indicator (CCI), part of the database on governability developed at the World Bank by the Kaufmann team.
See (Kaufmann, et al. 2006).
253
Table XI-1. Perception of Corruption in Seventeen Latin American countries

World
Regional CPI Score Surveys Range of
Country Ranking
Ranking (2006) used confidence
Canada 14 1 8.5 7 8.0 - 8.9
Chile 20 2 7.3 7 6.6 - 7.6
United States 20 2 7.3 8 6.6 - 7.8
Uruguay 28 5 6.4 5 5.9 - 7.0
Costa Rica 55 7 4.1 5 3.3 - 4.8
El Salvador 57 8 4 5 3.2 - 4.8
Colombia 59 9 3.9 7 3.5 - 4.7
Jamaica 61 10 3.7 5 3.4 - 4.0
Brazil 70 14 3.3 7 3.1 - 3.6
Mexico 70 14 3.3 7 3.1 - 3.4
Peru 70 14 3.3 5 2.8 - 3.8
Panama 84 18 3.1 5 2.8 - 3.3
Dominican Republic 99 21 2.8 5 2.4 - 3.2
Bolivia 105 22 2.7 6 2.4 - 3.0
Guatemala 111 23 2.6 5 2.3 - 3.0
Nicaragua 111 23 2.6 6 2.4 - 2.9
Paraguay 111 23 2.6 5 2.2 - 3.3
Guyana 121 26 2.5 5 2.2 - 2.6
Honduras 121 26 2.5 6 2.4 - 2.7
Ecuador 138 28 2.3 5 2.2 - 2.5
Venezuela 138 28 2.3 7 2.2 2.4
Haiti 163 30 1.8 3 1.7 - 1.8
Source: (Transparency International 2006a)
The CPI has been one of the best-known indexes available for making comparative
studies on corruption at a global level. However, it has also received criticism from different
sources, which has lead to a revision of its methodology. Part of the criticism concerns elements
inherent to the index, and, to a certain extent, question its validity. For instance, Seligson (2002,
2006) pointed out three weaknesses in the CPI and other measures of perception of corruption.
First, the index is based on perceptions about corruption and not the experience of corruption.
Second, the index is based on the opinions of national and international experts, which makes it
difficult to separate stereotypes from reality. Third, the measures of perception suffer from
endogenous problems, and therefore are not to be trusted as they can easily be influenced by the
media. Additionally, Luna (2006) argues that perceptions may be conditioned by the
circumstantial impact of specific scandals eventually becoming more volatile and becoming
strongly dependent of the moment in which the measure is taken (93).
Taking into account the weaknesses of the measurements of perceived corruption, the
Latin American Public Opinion Project (LAPOP) has developed a methodology that allows for
254
corruption measurement starting with direct citizen experiences of corrupt acts. The
AmericasBarometer included a battery of questions on corruption victimization that has
traditionally been used in LAPOP surveys. Respondents were asked about their personal
experience with corruption in different contexts: whether the police requested bribes; whether
they were asked to pay bribes to public employees; and whether they had to pay a bribe at work
or when they had dealings at the municipality, hospitals, schools and the courts. The following
table presents the LAPOP survey questions that are used to measure the degree of citizen
victimization by corruption.
Now we want to talk about your personal experience with things that happen in life...
EXC2. Did any police official ask you for bribe during the past year?
EXC11. During the past year did you have any official dealings in the municipality?
During the past year, to process any kind of document (like a license, for example), did you have to pay any money
above that required by law?
EXC14. During the past year, did you have any business in the courts?
Did you have to pay a bribe at the courts during the past year?
EXC15. Did you use the public health services during the past year?
255
Based on these items, LAPOP has created two corruption victimization indices. One takes
into account the total number of ways people have been victims of corruption over the past year;
the other one reflects whether the person has been a victim of corruption or not. Taking into
account the latter indicator, Figure XI-1 shows the percentage of persons that have been victims
of at least one act of corruption over the past year, per country.
Haiti 50.1%
Mexico 37.1%
Bolivia 34.6%
Jamaica 34.0%
Ecuador 31.9%
Peru 30.3%
Guyana 25.2%
Paraguay 21.7%
Venezuela 19.4%
Costa Rica 19.3%
Guatemala 18.0%
Nicaragua 18.0%
Dominican Rep. 17.7%
Honduras 16.1%
Brazil 15.6%
El Salvador 13.4%
Panama 11.3%
Colombia 9.7%
Chile 9.4%
Uruguay 7.0%
0 10 20 30 40 50 60

Figure XI-1. Percentage of the population victimized by corruption at least once in
past year
As can be seen in Figure XI-1, the LAPOP Corruption Victimization Index shows Chile
as having the lowest level of corruption among the countries studied. On the other hand, Haiti has
the worst indices of corruption victimization, which also seems to confirm Transparency
Internationals results. One out of every two Haitians has declared having been a victim of an act
of corruption, that is to say 50% of Haitians have had to pay bribes; this percentage is more than
256
double the regional average (22%).4 Although it may seem that the results of the Corruption
Victimization Index developed by LAPOP are similar to those of Transparency Internationals
CPI, there are substantial differences in the indicators. In fact, the correlation between the CPI
and the Corruption Victimization Index is relatively low (r=0.47). Whereas in Table XI-1 Costa
Rica appears to have one of the lowest perceived levels of corruption, according to LAPOPs
Corruption Victimization Index, this country has a level of corruption similar to Guatemala In
fact, in Central America, Guatemala, Honduras and Nicaragua had been placed among the
countries with generalized or alarming perceived levels of corruption but exhibit intermediate
levels of victimization. Moreover, Jamaica and Mexico at intermediate levels on the CPI become
part of the six countries where the magnitude of the corruption problem is serious, with a third of
their populations having been victims. Taking into account the aforementioned limitations of
corruption perception indexes, and the discrepancies between the results presented by the two
methodologies evaluated here, it can be concluded that LAPOPs Victimization by Corruption
Index allows for the measurement of the magnitude of corruption in Latin American countries in
a manner that is more in touch with reality, given that it measures peoples day to day experience
with acts of corruption.
Modes of Victimization by Corruption in Latin America and the Caribbean
An added strength of the methodology for measuring corruption developed by LAPOP is

that it not only allows for comparisons between countries but also reveals the contexts in which
acts of corruption take place. This is an important characteristic as it provides information that is
valuable for the design of national anti-corruption strategies since the places where acts of
corruption most frequently take place are identified. LAPOPs methodology focuses primarily on
measuring the levels of corruption victimization among public service users. In other words, to
calculate the percentage of persons who have been victims of corruption, only those who use
these services were taken into account. Figure XI-2 presents the percentage of persons in Latin
America and the Caribbean who have been victims of corruption, taking into account the place
where the person declares having been a victim. As can be seen, the levels of victimization by
corruption are greater in municipalities (15.2%) and at the courts (13%). Likewise, the data
indicates that, on average, only 6.8% of the people in the region reported having been a bribe-
victim of a public employee.
4
For more information on corruption in Haiti, see (Zphyr, et al. 2007).
257
20%
15%
Per cent Victimized
10%
15.2
13.0
5% 9.8 9.9 10.1
9.0
6.8
0%
Public Health Bribery at School Police Court Municipal
Employee Service Work Bribery Bribery Bribery Bribery
Bribery Bribery

Figure XI-2. Modes of victimization by corruption in Latin America and the Caribbean
These regional results hide substantial differences among countries. For instance, results
per country for the question Have you been asked to pay a bribe by a public official in the past
year? are presented in Figure XI-3. This is the only item in the battery of questions on
corruption victimization that was included in the questionnaires used in Canada and the United
States. This allows for the comparison of levels of victimization by public employees in Latin
America with those in the two countries mentioned. As can be seen, Ecuador, Bolivia, Mexico
and Paraguay appear at the top of the graph with relatively homogenous levels of corruption
victimization. Citizens in these countries are 40 times more likely to be asked for a bribe by a
public official than people in the U.S. On the other hand, Chile, which has the lowest level of
victimization among Latin American countries, has twice the percentage of people who reported
having paid a bribe to a public employee than Canada does.
258
Ecuador 15.1
Bolivia 14.5
Mexico 13.7
Paraguay 13.1
Haiti 10.8
Peru 10.7
Venezuela 9.4
Guyana 6.4
Dominican Rep. 6.3
Costa Rica 6.1
Guatemala 4.7
Nicaragua 4.1
Honduras 3.9
Panama 3.7
Colombia 2.6
Brazil 2.6
El Salvador 2.5
Jamaica 2.3
Uruguay 1.8
Chile 1.6
Canada 0.7
United States 0.3
0% 5% 10% 15% 20%

Figure XI-3. Public Employee requested bribe
In Latin America and the Caribbean, the highest levels of victimization by corruption can
be found in the municipalities. Some 15.2% of people have been victims of corruption when they
had dealings with these institutions. This shows the importance local government has in the life
of the Latin American citizen, but it also demonstrates that corruption is a serious problem for
local governments. As shown in Figure XI-4, the problem of corruption in municipalities is very
serious, particularly in Haiti and also in Bolivia, Mexico, Brazil, Dominican Republic, Jamaica
and Panama where percentages of people who have had to pay bribes are higher than the regional
average (16%). Corruption in municipalities has been mentioned as one of the three most
frequent modes in all the countries in the region that were included in the AmericasBarometer.
259
Haiti 62.0
Bolivia 24.2
Mexico 24.1
Brazil 22.9
Dominican Rep. 19.3
Jamaica 16.0
Panama 15.9
Peru 14.9
Ecuador 14.6
Venezuela 14.2
Guyana 13.6
Paraguay 12.9
Nicaragua 12.6
Honduras 10.6
Guatemala 6.4
El Salvador 6.0
Costa Rica 5.9
Chile 5.6
Colombia 4.4
Uruguay 1.8
0 20 40 60

Figure XI-4. Bribes in municipalities
This result turns out to be of particular relevance in the Latin American context given that
decentralization of the central governments resources towards municipalities is being promoted
as a strategy to foster democratic governance, as it is believed that this will lead to greater citizen
participation and therefore more accountability of public officials. However, as shown in Chapter
V of this book Decentralize or Centralize? Challenges for Reform of the State and Democracy
in Latin America and the Caribbean, decentralization does not have general citizen approval.
Actually, only one of every two Latin American citizens supports decentralization. A hypothesis
that emerges when taking into account the results in this chapter is that the divided opinion on
decentralization in Latin America may be related to the high rates of corruption experienced at
the municipal level. Figure XI-5 shows evidence in favor of this hypothesis, given that those who
have had to pay bribes to municipal government officials show a lower level of support for
decentralization of resources and powers toward local governments. The results of the
AmericasBarometer suggest that decentralization must go hand-in-hand with efforts to fight
corruption in municipalities so that decentralization will be supported by the majority of the
260
population and also in order to avoid an increase in the levels of victimization by corruption
provoked by transferences of power and resources to the municipalities.
60.0
50.0
Support for decentralization
40.0
30.0
53.4
46.7
20.0
10.0
0.0
Not-victim Victim
Source: AmericasBarometer 2006 Error Bars:95% CI

Figure XI-5. Impact of victimization by corruption at municipalities on support for decentralization
The second highest level of victimization takes place at the courts. On average, some 13%
of respondents claimed to have been victims of corruption when they had dealings at the courts.
Figure XI-6 indicates that approximately 3% of people who had some business at the courts were
asked to pay bribes in El Salvador, Costa Rica and Colombia whereas one out of every two
Haitians and one out of every four Mexicans had an experience of bribe payment at the courts.
261
Haiti 50.2
Mexico 25.0
Ecuador 23.2
Nicaragua 22.9
Bolivia 18.9
Paraguay 16.9
Jamaica 16.9
Panama 14.1
Dominican Rep. 12.4

Peru 11.6
Guyana 10.2
Honduras 7.4
Venezuela 6.8
Brazil 6.5
Guatemala 6.3
Chile 5.2
Colombia 3.3
Costa Rica 3.0
El Salvador 2.4
Uruguay 0.0
0 10 20 30 40 50 60

Figure XI-6. Bribes at court
As shown in Chapter XII of this book, being a victim of corruption has a negative impact
on trust in the institutions that guarantee the rule of law in democratic Latin American societies,
and the high level of corruption in the courts seems to partially explain this result. As shown in
Figure XI-7, in fact, persons that have been victims of corruption at the courts have a lower level
of trust in the justice system.
262
50
Trust in the Justice System
40
30
45.5
20 39.1
10
0
Not-victim Victim

Figure XI-7. The impact of victimization by corruption at courts on levels
of confidence in the justice system
Who are the most likely to become victims of corruption?
In an effort to identify potential victims, a logistic regression model was made. In this
model the dependent variable indicates whether or not the respondent was a victim of at least one
act of corruption over the past year. The independent variables are gender, age, educational level,
income level, marital status and the size of the city of residence. These variables have been taken
into account given that previous analyses have consistently shown that corruption does not affect
all citizens equally.5 The regression results are shown in Table XI-2. It can be seen that with the
exception of marital status, the majority of variables that were included are statistically relevant
factors (p<0.05). Based on the coefficients estimated in the regression, the probability of
becoming a victim was calculated taking into account the interviewees gender and maintaining
other variables constant at their average value. Furthermore, the probability of becoming a victim
was calculated taking into account the individual effect of age, income and educational level.
5
See studies of the results from the LAPOP national surveys at www.lapopsurveys.org
263
Table XI-2. Determinants of victimization by corruption in Latin America and the Caribbean
(Results of the logistic regression)
Coefficient Standard Dev. t Sig.
Gender 0.490 0.033 15.040 0.000
Age -0.012 0.001 -8.290 0.000
Education 0.059 0.006 10.680 0.000
Income 0.019 0.010 1.830 0.067
Marital status (married) 0.085 0.037 2.300 0.022
Number of children 0.070 0.010 6.950 0.000
Size of city of residence -0.017 0.019 -0.900 0.369
Constant -1.880 0.111 -16.870 0.000
Dependent Variable: 0= Not victim; 1= Victim of corruption
Number of observations: 25,289
Effect of design = 1581
F(7, 1575) = 71.68
Figure XI-8 shows that women have a lower probability of becoming victims of
corruption than men. On average, women are 8% less probable to become victims of corruption.
This is due to the fact that, in Latin America, there is still a gap in the level of participation in
public affairs between men and women, to the detriment of the latter. This reduces the number of
situations in which women could be victimized.
264
25%
Probability of being victimized by

corruption 20%
15%
25.0
10%
16.9
5%
0%
Man Woman
Figure XI-8. Probability of becoming a victim of corruption by gender
As shown in Figure XI-9, young people are more prone to becoming victims of
corruption. Considering that municipalities are the institutions where there are higher levels of
corruption, it can be inferred that young people have a greater tendency to go to these institutions
to do tasks such as request birth certificates or permits, which increases the opportunities of being
exposed to a greater level of victimization by corruption as compared to older people.
265
30.0%
Probability of being victimized by corruption
25.0%
Probability of being victimized by

25.0%
22.5%
corruption
20.0%
20.0%
17.5% 15.0%
15.0%
10.0%
None Primary Secondary Higher

education
16-25 26-35 36-45 46-55 56-65 66+
Source: AmericasBarometer 2006 Source: AmericasBarometer 2006
Figure XI-9. Probability of becoming a victim of Figure XI-10. Probability of becoming a victim of
corruption by age corruption by educational level
On the other hand, people with a higher socioeconomic status are expected to have the
highest tendency to become victims of corruption, given that they have greater economic
resources and are therefore more prepared to pay a bribe. They also participate more in public
affairs. Figure XI-10 shows that education is a factor that is associated with victimization; the
greater the level of education, the greater the probability of becoming a victim of corruption.
Furthermore, the regression results indicate that there exists a positive, statistically significant
relationship between income and corruption victimization. These results confirm the notion that
persons with a higher socioeconomic level have greater probabilities of becoming victims of
officials demanding bribes.
Corruption and its Impact on Democracy
This section examines the impact corruption has on democratic stability. The
development of a democratic culture among citizens, and in particular the level of support or
preference for democracy, is recognized in the literature as an important element for achieving
durable and stable democracies (Almond and Verba 1963; Diamond 1999). This section analyzes
the degree to which incidence of demands for bribes by public employees affects the opinions of
266
citizens regarding democracy. The study specifically focuses on whether being a victim of
corruption affects support for democracy as it exists, measured by the degree of individual
satisfaction with the way in which democracy functions in the countries studied and the support
for democracy as an ideal form of government, that is--normative support for democracy.
Likewise, this section analyzes the effect of victimization by corruption on citizen

opinions regarding the legitimacy of the political system. Easton (1965; Easton 1975) defined
political legitimacy as the conviction that it is right and properto obey the authorities and to
abide by the requirements of the regime (1975, 451). Similarly, Lipset (1994) states that
legitimacy involves the capacity of a political system to engender and maintain the belief that
existing political institutions are the most appropriate or proper ones for the society and
considers legitimacy as a prerequisite for the maintenance and the stability of democracy. The
argument is that a political system that is legitimate in the citizens view will obtain a high level
of support for its institutions. That is why, aside from examining the effect of being a victim of
corruption on the satisfaction and support for democracy, its effect on political legitimacy is also
studied. It is assumed that corruption victimization in Latin American countries will decrease the
political systems legitimacy and that at the same time it may lead to less support for democracy,
given that corruption causes citizens to lose trust in the political systems institutions and the
performance of democratic regimes.
Victimization by Corruption and Satisfaction with Democracy
The first dependent variable analyzed is satisfaction with the way democracy works. In
the AmericasBarometer, this variable is measured using the following question: In general,
would you say you are very satisfied, satisfied, dissatisfied or very dissatisfied with the way in
which democracy functions in [country]? Satisfaction with democracy reflects citizen perception
of the way the democratic regime actually works in their country. That is to say, satisfaction
refers to the way democracy actually functions, not democracy as a form of ideal government in
itself (Bratton, et al. 2004).
Figure XI-12 shows that the level of satisfaction with the way democracy works in Latin
America is relatively low. The regional average for satisfaction with democracy is 46.8 on a scale
of zero to one hundred points. Nevertheless, satisfaction with democracy varies from country to
country. As can be seen in the Figure, countries with the lowest levels of satisfaction with
democracy are Paraguay (27.6) and Ecuador (36.5). On the other hand, Uruguay exhibits the
highest level of satisfaction with democracy (62.2), followed by Bolivia (54.1).
267
Uruguay 62.2
Bolivia 54.1
Dominican Rep. 53.3
Venezuela 53.3
Chile 52.5
Costa Rica 52.4
Colombia 51.6
Jamaica 51.5
Honduras 50.2
Mexico 48.7
Panama 46.9
Guyana 46.5
El Salvador 46.4
Nicaragua 45.7
Guatemala 43.6
Brazil 43.6
Peru 42.9
Haiti 37
Ecuador 36.5
Paraguay 27.5
0 20 40 60

Figure XI-12. Satisfaction with the way democracy works
In order to determine the impact of victimization by corruption on satisfaction with the

way democracy works, a regression model was constructed. The predictors included in the
regression are the following: (a) the total number of ways that people have been victims of
corruption over the past year; (b) the usual sociodemographic variables: gender, age, education,
wealth, and place of residence; and (c) dummy variables or dichotomic variables for each country
which take into account the differences in the levels of satisfaction with democracy from country
to country. In addition, variables have been included to measure the level of satisfaction with the
performance of the incumbent government, the perception of the countrys economic situation,
and victimization by crime. The results of the regression are presented in Table XI-3. These
indicate that the experience of corruption influences people negatively when assessing the way
democracy functions.
268
Table XI-3. Determinants of satisfaction with the way democracy works

Coefficient. Stand.Dev. T Sig.
Total number of ways they have been victims of
corruption over the past year -0.534 0.165 -3.240 0.001
Satisfaction with the performance of the
current president 0.297 0.006 49.720 0.000
Victimization by crime -0.023 0.003 -6.610 0.000
Perception of the countrys
economic situation 0.155 0.006 25.090 0.000
Gender 0.158 0.255 0.620 0.535
Age 0.024 0.009 2.710 0.007
Education -0.772 0.192 -4.020 0.000
Wealth -0.254 0.080 -3.160 0.002
Urban -1.249 0.300 -4.160 0.000
Constant 29.988 0.882 33.990 0.000
Number of obs =29469
F(28, 29440) = 278.01
Adj R-Squared. = 0.2084
Dependent Variable: Satisfaction with the way democracy works.
Dummy for countries (Reference country: Mexico).
Victimization by Corruption and Support for Democracy as an Ideal Form of

Government
As mentioned earlier, one of the prerequisites for the consolidation of democracy in the
country is generalized citizen support for democracy as an ideal form of government. In the
AmericasBarometer, the level of support for democracy is measured with the following question,
using a scale from one to seven: Democracy may have problems but it is better than any other
form of government. To what extent do you agree or disagree with this statement?
According to Figure XI-13 the level of support for democracy as the best form of
government is high in all the countries of the region. The regional average for support of
democracy is 69.9. In fact, Peru and Panama, countries with the lowest levels of support for
democracy, have a level of over 50 points on the scale from zero to 100. These figures suggest
that the demand for democracy is greater than the supply, this is to say, the citizens in the region
have a high level of support for democracy as the ideal form of government, whereas they have
low levels of satisfaction on how democracy actually works in their countries.
269
100
Demand ( Support for Democracy)

80
60
40
Supply (Satisfaction with Democracy)
20
U
D uay
V nic
C zu
Ja
H ia
G ura
M na
B co
G l
H em
El
N l va
Pe ag
Ec
P a do
om
en
os el ep
hi ca
ol
ol bia
ra
ru
on
uy
ua
ai
ic
ex
m ica
Sa
ru ua
ua
na r
le
om
iv
ti
ar dor
zi
g
ta a
e
a
d
t
i
ai
i
m
R
a
al
an
a
R
ub
lic
Figure XI-13. Supply and demand for democracy in Latin America and the Caribbean
Does victimization by corruption affect support for democracy as the ideal form of
government? In order to answer this question, the previous regression model was estimated, this
time with support for democracy as the dependent variable. The results (see Table XI-4) showed
that corruption victimization has a statistically significant negative effect on Latin American
citizens propensity for expressing support for democracy as a form of government. This result
contradicts Canache and Allisons conclusions (2003), which were based on measures of
perception of corruption. They had found no empirical evidence that corruption affects citizens
opinions regarding democracy as a form of government.
270
Table XI-4. Determinants of support for democracy as a form of government

Coefficient Stand. Dev. T Sig.
Satisfaction with the performance
of the current president 0.148 0.008 18.890 0.000
Gender 0.896 0.330 2.710 0.007
Age 0.148 0.012 12.830 0.000
Education 2.703 0.248 10.910 0.000
Wealth 0.430 0.104 4.150 0.000
Urban -0.831 0.389 -2.130 0.033
Constant 48.234 1.136 42.470 0.000
Number of obs =28261
F(27, 28233) = 105.12
Adj. R- Squared = 0.0905
Dependent Variable: Support for democracy as a form of government
Dummy for countries (Reference country: Mexico)
Victimization by Corruption and Support for the Political System
The AmericasBarometer includes a battery of five questions that use a scale from one to
seven to measure the political systems legitimacy among citizens6. The results of the regression
on support for the system are presented in the Appendixes. The variable for the study,
victimization by corruption, has a negative relation to legitimacy for the system, even after the
control variables are taken into account, as can be seen in Figure XI-14. The figure shows that
interviewees who were not victims of corruption have a significantly greater level of confidence
in the system. As the number of acts of corruption increase, their support for the system
diminishes. Furthermore, this result is supported when using LAPOPs index of victimization by
corruption. In conclusion, these results confirm the empirical evidence from previous studies
which found that corruption erodes support for the political system and therefore decreases
citizen perception of its legitimacy (Seligson 2002, 2006; Canache y Allison 2005).
6
B1. To what extent do you think the courts of justice guarantee a fair trial?
B2. To what extent do you respect the political institutions of (country)?
B3. To what extent do you think that citizens basic rights are well protected by the political system of (country)?
B4. To what extent do you feel proud of living under the political system of (country)?
B6. To what extent do you think that one should support the political system of (country)?
271
52
System Support
50
48
45
42
0 1 2 3 4+
Total number of ways victimized by corruption in last year
Figure XI-14. The Impact of Victimization by Corruption on Support for the System
Conclusions
This study used the methodology developed by the Latin American Public Opinion
Project to determine the magnitude of the problem of corruption in Latin America and the
Caribbean. This methodology measures the level of corruption based on actual citizen experience
with bribery, instead of their perception of corruption. One of the advantages of this method is
that it shows who are most likely to become victims of corruption and where acts of corruption
occur. In general, the results of the analysis, which was based on AmericasBarometer data,
confirm the fact that corruption is a serious problem in the region.
The data show that municipalities and the courts are the principal sites for corrupt
practices. Furthermore, it was found that becoming a victim of corruption negatively affects
citizen trust in the justice system and citizen support for the decentralization of resources from
the central government to local governments. This suggests that success for programs of
decentralization and the strengthening the rule of law in Latin America and the Caribbean
depends to a great extent on inroads made in the fight against corruption in municipalities and in
272
the justice system. Additionally, it was found that, just as previous LAPOP studies suggest,
corruption does not affect everyone equally. It has been seen that men, people of higher economic
status, and the young are the most likely to be the victims of corrupt acts.
This study has found that corruption constitutes an obstacle for the consolidation of
democracy in the region, given that empirical evidence has found that victimization by corruption
reduces the level of support for democracy as an ideal form of government and decreases citizen
satisfaction with the democratic regimes performance. Furthermore, evidence indicates that
corruption degrades the level of confidence in state institutions, therefore diminishing the
political systems legitimacy.
A positive note is that this chapter highlights the fact that Latin American citizens see
democracy as an ideal form of government, in spite of their dissatisfaction with the manner in
which it functions in their respective countries. However, it was also shown that corruption can
have harmful effects on the supply and demand of democracy. Should the current patterns
continue, in the long run, democratic stability in the region may be undermined, principally in
countries with higher levels of corruption.
273
References
Almond, Gabriel A., and Sidney Verba. The Civic Culture: Political Attitudes and Democracy in
Becquart-Leclercq, Jeanne. "Paradoxes of Political Corruption: A French View." In Political
Corruption: A Handbook, edited by Arnold J. Heidenheimer, Michael Johnston, Victor
T. LeVine, 191-210. New Brunswick, NJ, 1989.
Bratton, Michael, Robert B. Mattes, and Emmanuel Gyimah-Boadi. Public opinion, democracy,
and market reform in Africa, Cambridge studies in comparative politics. New York:
Camp, Roderic Ai, Kenneth M. Coleman, and Charles L. Davis. "Public Opinion About
Corruption: An Exploratory Study in Chile, Costa Rica and Mexico." Portland, Oregon,
2000.
Canache, Damarys, and Michael E. Allison. "Perceptions of Political Corruption in Latin
American Democracies." Latin American Politics and Society 47, no. 3 (2005): 91.
Della Porta, Donatella. "Social Capital, Beliefs in Government, and Political Corruption." In
Disaffected Democracies: What's Troubling the Trilateral Countries?, edited by Susan J.
Pharr and Robert D. Putnam, 202-28. Princeton: Princeton University Press, 2000.
Easton, David. A Systems Analysis of Political Life. New York,: Wiley, 1965.
. "A Re-Assessment of the Concept of Political Support." British Journal of Political
Science 5 (1975): 435-57.
Huntington, Samuel P. Political Order in Changing Societies. New Haven: Yale University
Press, 1968.
Kaufmann, Daniel, Aart Kraay, and Massimo Mastruzzi. "Governance Matters V: Aggregate and
Individual Governance Indicators for 1996-2005." (2006).
Lipset, Seymour Martin. "The Social Requisites of Democracy Revisited." American
Sociological Review 59 (1994): 1-22.
Luna, Juan Pablo. Cultura poltica de la democracia en Chile: 2006. Edited by Latin American
Public Opinion Project. Nashville, TN: Vanderbilt University, 2006.
Pharr, Susan J. "Officials' Misconduct and Public Distrust: Japan and the Trilateral
Democracies." In Dissaffected Democracies: What's Troubling the Trilateral Countries?,
edited by Susan J. Pharr and Robert D. Putnam. Princeton: Princeton University Press,
2000.
Seligson, Mitchell A. "The Impact of Corruption on Regime Legitimacy: A Comparative Study
of Four Latin American Countries." Journal of Politics 64 (2002): 408-33.
. "The Measurement and Impact of Corruption Victimization: Survey Evidence from
Latin America." World Development 34, no. 2 (2006): 381-404.
Shin, Doh C. Mass Politics and Culture in Democratizing Korea. Cambridge: Cambridge
Transparency International. "The Corruption Perception Index." Transparency International,
2006a.
. "CPI 2006 Regional Results: Americas ": Transparency International, 2006b.
274
Waterbury, John. "Corruption, Political Stability, and Development: Comparative Evidence from
Egypt and Morocco." Government and Opposition 11 (1976): 426-45.
World Bank. World Development Report, 1997. Washington, D. C.: Oxford University Press,
1997.
Zphyr, Dominique, Yves-Franois Pierre, and Abby Crdova. Culture politique de la
dmocratie en Hati: 2006. Edited by Latin American Public Opinion Project. Nashville,
TN: Vanderbilt University, 2007.
275
Appendix
Table XI-5. Determinants of support for the system

Coefficient. Stand. Dev. t Sig.
Satisfaction with the performance of
the current president 0.258 0.006 45.830 0.000
Gender -0.488 0.240 -2.040 0.041
Age 0.042 0.008 4.940 0.000
Education -0.187 0.181 -1.040 0.299
Wealth -0.076 0.075 -1.010 0.314
Urban -1.475 0.282 -5.240 0.000
Constant -1.408 0.155 -9.070 0.000
Number of observations = 30225
F(28, 30196) = 285.71
Adj. R-squared = 0.2087
Dependent Variable: Index of support for the political system.
Dummy for countries (Reference country: Mexico).
276
XII. Justice and democracy: the Rule of Law in the Americas

Juan Carlos Donoso1
Abstract
The administration of justice is based on the ability of the judicial system to give the citizens what
they are entitled to. Judicial independence is an indispensable requisite for the proper
functioning of judicial power in the Americas. This chapter explores the effects that structural
variables, such as the autonomy of the justice system have on personal attitudes of citizens
towards that system and the consequences of those attitudes in the possibilities for the
consolidation of liberal democracy in the continent.
Judicial reform and the consolidation of the autonomy of the justice system have been at
the top of the list in Latin America when it came time to design new public policies. These
reforms have been aimed to strengthen the judicial power with respect to the Executive and the
Legislature, but also to improve the way the case load is managed, the competence of the judges
and the overall mechanisms of conflict resolution. The relevance of justice reform, however,
may seem diminished when the system overall lacks legitimacy among the population. The
present chapter takes on the subject of the legitimacy of the justice system and citizen support for
the institutions that make out the rule of law, especially in Latin American countries, as well as
the consequences that these attitudes may come to have on the regions democratic consolidation.
The first section of the chapter will focus on the effects that judicial strength has on individual
attitudes towards the judicial system. A second segment will look at individual level variables
that act as predictors of citizen support for the judiciary. Finally, I will examine the
consequences that a low degree of support for the institutions of the rule of law can have on the
process of democratic consolidation in Latin America, as well as on the raise of new hybrid
democracies in the region.
Judicial Power
The strength of the justice system in a country is a key factor for the legitimacy of a
democratic regime. The consolidation of democracy can only take place once a series of
requisites have been fulfilled, one of which is the formation of an autonomous justice system that
can generate in its citizens a feeling of trust in the rule of law (Becker 1999, Carothers 2006). A
democracy with a weak judicial system may have serious problems in securing popular support
for the rule of law (Prillaman 2000, Verner 1984). Holston and Caldeira (1998) observe that the
concept of citizenship, indispensable in a democracy, includes as a minimum the notion of
justice, legality, access and universality of the law. When the majority of the population
perceives the justice system as inaccessible and inefficient, this may foster a feeling of
1
I would like to thank USAID for the funding provided to LAPOP and the AmericasBarometer 2006. My thanks
especially to the USAID mission in Ecuador, which has funded my studies at Vanderbilt University and has
supported my academic career at all times.
277
incomplete citizenship, which makes their commitment to the regime decrease in intensity or
disappear all together (O'Donell 1994).
The strength of the justice system in a country can be interpreted in different ways. Some
of the existing literature which deals with the subject identifies three main characteristics of a
healthy judiciary. The first one is the free access of the citizenry, meaning that everyone should
have the same opportunity to utilize the services of the justice administration system. For this to
happen, two conditions must be fulfilled; physical presence of the judicial institutions throughout
the territory and reasonable costs to those who demand the services of the justice system, so that
everyone can use it. The latter condition is linked to the second characteristic of a healthy
judiciary, which is efficiency. An effective judiciary must comply with two basic requisites.
First, to process all cases within the time frames established by law and second, to execute the
sentences dictated by the appropriate judges. The third and most important characteristic of a
strong judiciary is the autonomy of the judicial system with respect to the executive and the
legislature and the independence to rule against them (Cappelletti 1985). Although the terms
autonomy and independence are similar, one must separate one from the other theoretically.
Autonomy is an institutional trait that allows the judicial system to build its own institutional
structure (Rios-Figueroa 2006). Independence, on the other hand, stems from the principle of
conflict resolution by a neutral party (Larkins 1996). An independent judiciary, then, is one that
can freely decide, based on the law and the facts and without taking into consideration the
preferences of the parties involved (Fiss 1993).
Even though the constitutions of every Latin American country guarantee the autonomy
of the judicial system, these provisions are not always applied in practice due to several reasons,
such as power grabbing by the executive, the constant interference of political interest groups or
pressure from public opinion in certain matters. In this study, judicial strength will be
determined based on the annual reports from the U.S. State Department (www.state.gov). These
reports, which begun in 1977, are published annually and their goal is to describe the efforts that
every country makes to put in practice their commitment to protect the human rights of its
citizens. The reports are individual and contain a detailed description of human rights violations
that were reported in each country by affected individuals or by national or international media
outlets. These violations are then analyzed in different contexts, on e of which is due process.
Within this analysis, the reports evaluate the strength of the judicial system of every country and
the main problems that they face. Although these chronicles are subject to the criteria of their
authors, this is one of the few systematic sources of information about judiciaries in Latin
America. Various academic studies written about the judicial system in the region ratify the
validity of these reports (Cross 1999, Dakolias 1994, Howard and Carey 2004, Yamanashi 2002).
278
The Data
This section will attempt to demonstrate the importance of the strength of a judiciary has
on the trust that the citizens of a country have in it. Conficence in the justice system in the
Americas is measured based on the data from the 2006-2007 AmericasBarometer, gathered in 19
countries in the region. In Latin America and the Caribbean, the interviews were conducted face
to face, while in the United States and Canada phone surveys were carried out and a shorter
version of the questionnaire was used due to budget constraints. The questionnaire was produced
by the Latin American Public Opinion Project LAPOP- which is based at Vanderbilt University.
To measure the legitimacy of the justice system, the AmericasBarometer utilizes the following
question.
B10A. To what extent do you trust the justice system?
In order to answer this question, the respondent is shown a scale from 1 to 7 in which 1
means not al all and seven means a lot. However, to facilitate the readers comprehension,
every scale has been turned into a 0 to 100 interval. Figure XI.1 illustrates how the countries in
the Americas are distributed according to their level of confidence in the justice system. It is
notable that the mean of the whole region is below the midpoint of the scale, at a score of only
42. Canada is the country where citizens trust the judiciary the most with a score of 64 in the
scale from 0 to 100. The United States and Costa Rica follow closely, which is not a surprise, if
one considers that those are thought to be the three strongest democracies in the region. The
presence of Guyana, a small and relatively unknown country among the nations with the highest
scores is unexpected. It could be speculated that the British influence in Guyana is at least
partially responsible for the high level of support for the judiciary; however this is not a tendency
that repeats itself in other countries that share that heritage, such as Jamaica, which finds itself
just below the mean of the region, as the Figure shows. At the lower end of the scale is Ecuador,
with a score of 28, less than half than Canada and well below the average for the Americas.
Considering this countrys recent political history, however, its levels of trust with compared to
the other countries of the region is not all that surprising.
279
60
Mean Trust in The Americas
40
63.8
61.8
55.6
52.9 51.7
50.7 50.6
48.2 46.7 46.6
46.0 45.3 44.7 44.6
43.4 42.0
41.2 40.9
20 39.5
32.6 31.0
28.0
0
C
U da
C na
El
D lva
V nic
G zue Re
H em
Pa
Ja ma
C ia
Pe
Pa
Ec ua
an
os y
ol
om d
en
ra
ol
hi
ni
ru St
uy Ric
ua
on
ic
ai gu
ex ia
m
Sa
na s
ru
ra
ua y
om
iv
le
te
ar
ti
si
gu at
ta
a
du a
ic
ai
i
do
l
d
a
a
ca
b
ra
r
al
an
la pu
a
or
a
es
bl
ic
Source: AmericasBarometer 2006 by LAPOP Error Bars: 95% CI
Figure XII-1. Level of Confidence in the Justice System in the Americas
Which factors explain the large differences that exist between countries? A fraction of the
answer lies in the experiences, circumstances and attitudes of the citizens of each country. The
rest of it, however, goes beyond individual differences. Each bar of Figure XII-1 shows the level
of confidence of an average citizen in a given country. An analysis of variance suggests,
however, that at least 15% of the variation in the dependent variable level of confidence in the
justice system- can be explained by difference between countries and not between individuals.
This means that structural characteristics, like judicial strength in this case, can mold the attitudes
of individuals (Seligson and Muller 1994).
As I mentioned above, the strength of the judicial system in the countries of the region
was measured by looking at the individual Human Rights Country Reports published yearly by
the U.S. State Department. Since these reports only give a descriptive narration of the strengths
and weaknesses of the judicial system in each country, it was necessary to devise a coding
scheme in order to systematize the information. To this order, a three point scale was built, in
which a score of one would be assigned to countries with low levels of autonomy and judicial
independence. Countries with mid and high levels of autonomy and independence received a
280
score of two and three respectively. I based the scoring criteria on the one used previously by
Cingranelli and Richards as well as the one used by Tate, Poe and Keith to measure the degree of
human rights protection in a country (Tate, et al. 1999, www.humanrightsdata.org). This was the
criteria used for assigning individual scores to each country:
1.- Low: The judiciary is not autonomous or independent. The interference from the executive is
constant and judges can be removed without warning.
2.- Medium: The judiciary is described as partially independent. Executive influence is less
notorious but judges do not enjoy full protections.
3.- High: The judiciary is generally independent. The courts can rule against the executive
without fear of repercussions and judges enjoy full constitutional protections.
Figure XII-2 shows the scores that each country in the sample was assigned. Only four
countries, Canada, The United States, Costa Rica and Chile, received a score of three, while eight
countries, nearly half the sample, received a score of one. The values assigned to the countries
according to this scale mostly coincide with the levels of confidence shown by the citizens of
those countries in their justice systems, as shown in the previous graph.
To capture the impact that a structural trait such as the strength the judiciary of a country
can have in the individual level of confidence that a citizen has on his justice system I have
developed a hierarchical lineal regression model, which measures the effect that judicial strength
has on individual confidence, controlling for other individual predictors that may also influence
the dependent variable. These predictors include the level of education, the perception of
corruption in the public sector, satisfaction with democracy and victimization by violence and
crime during the year previous to the survey. Table XII-1 shows the result of the regression
equation.
281
United States 3.0
Canada 3.0
Costa Rica 3.0
Uruguay 3.0
Chile 3.0
Colombia 2.0
Jamaica 2.0
Peru 2.0
Brasil 2.0
Mexico 2.0
Guyana 2.0
Panama 2.0
Paraguay 1.0
Honduras 1.0
Ecuador 1.0
Venezuela 1.0
Nicaragua 1.0
Bolivia 1.0
Guatemala 1.0
Haiti 1.0
El Salvador 1.0
0.0 1.0 2.0 3.0
Judicial independence
Figure XII-2. Judicial Strength measured on a scale from 1 to 3.
282
Table XII-1. Effect of Judicial Strength on Confidence on Judicial System

Coef. Std. Err. P>z
Judicial Strength 0.242 0.110 0.028

Education -0.048 0.016 0.002
Perception of corruption 0.181 0.013 0.000
Crime victimization 0.212 0.029 0.000
Satisfaction with 0.530 0.016 0.000
democracy
Income -0.005 0.006 0.382
Constant 4.068 0.232 0.000
As is shown in the table, the effect of judicial strength on individual trust in the judiciary
is positive and significant. This means that the more autonomous and independent the judiciary
of a country is, the more the citizens of that country will trust it, as is illustrated on Figure XII-3.
Other variables that were also significant are perception of corruption and having been victimized
by violent crime. The regression equation shows that people who perceive a high degree of
corruption among public officials as well as citizens who were victimized by crime show lower
levels of trust in their justice systems. Finally, education is another factor that has an important
effect, although the relationship between this variable and trust in the judiciary is negative,
meaning that people with higher levels of education tend to have lower levels of trust in the
justice system than those with low levels of education.
283
52.5
50
47.5
45
42.5
40
Low Medium High
Judicial Independence
Figure XII-3. Effect of Judicial Independence on Individual Confidence on Justice System
It is important to note that the effect of judicial strength on individual confidence on the
justice system could be attenuated by other predictors that were not included on the regression
equation. The reason behind this omission is that some questions were not included in the
questionnaires used in the United States and Canada, since brevity has to be prioritized when
conducting a survey via telephone. The results shown in this model, whoever, show that even
controlling for other individual level variables, such as personal experiences and perceptions;
structural factors do create personal attitudes.
The following section will focus on uncovering the predictors of individual support for
the institutions of the rule of law in Latin America. Now that I have shown that structural
variables can have consequences in the generation of individual attitudes, the next pages will
emphasize some of the individual level predictors that shape those attitudes. It is important to
284
mention that Canada and the United States will not be included in that analysis due to the
limitations that I have already addressed.
Support for the Institution of the Rule of Law
While most nations in Latin America have undergone a transition process to electoral
democracy, many of the legal institutions in these countries still maintain authoritarian traits.
This is a problem which is not exclusive to Latin America, but to the majority of democracies
born from what is now known as the third wave of democratization (Huntington 1991, O'Donnell
and Schmitter 1986). The path that these democratic regimes have walked clearly shows that the
institutionalization of an electoral process is achieved more easily than that of the rule of law
(Pereira 2000, Schedler, et al. 1999). But what is the rule of law? It is generally understood that
the rule of law is intimately connected to the capacity of a legal system to apply the law
efficiently and equally (Seligson 2004), but the formulation of a precise definition is a
complicated job due mostly to the multiple dimensions of the concept (Goertz 2006). Many
theoretical discussions have taken place about what constitutes a definition of the rule of law and
what are its necessary and sufficient conditions (Collier, et al. 2006)
The current section has the goal of finding the variables that determine individual citizen
support for the institutions that make up the rule of law in Latin America and the Caribbean. To
this effect, I have chosen a battery of questions that attempt to capture some of the dimensions of
the rule of law. It is important to note that the group of questions selected by no means covers
every dimension of this essentially contested and complex concept. However, it is a good
practical approximation to the effect of the analysis that will be done in this chapter. The
questions in the AmericasBarometer that measure support for the institutions of the rule of law
are:
B1. To what extent do you believe that you can get a fair trial under this countrys legal system?
B10A. To what extent do you trust the justice system?
B18. To what extent do you trust the police?
B31. To what extent do you trust this countrys Supreme Court?
As I stated above, these questions contain some of the dimensions of the concept of rule
of law, such as the necessity for due process, the institutionalization of the justice system and the
states duty to provide security for its citizens (Schor 2003). The questions have been combined
to create an index of citizen support for the institutions of the rule of law. A reliability test
indicates that the indexs Cronbachs Alpha is of .756, which means that the questions selected
for the index of citizen support for the institutions of the rule of law are highly correlated and do
not measure different concepts. A factor analysis also shows that the four variables create a single
dimension. Figure XII-4 shows how the countries of Latin America and the Caribbean are
distributed according to the level of their citizens trust for the institutions that make up the rule
of law. As the graph shows, there is an average score of 45 on a scale from 0 to 100 for the
whole region, which again is worrisome, since it does not reach the midpoint of the scale. Again
there are big differences visible between countries. Those countries at the top of the scale, such
285
as Guyana, Colombia and Costa Rica, score over 20 points higher that the countries at the bottom
of the scale, like Ecuador and Paraguay.
Mean Trust in the Institutions of the Rule of
60.0
Mean Trust in the Americas
40.0
Law
67.2
63.3
55.1
53.2 52.4
50.4 49.8 49.2 48.8 48.6 47.6 46.7 46.3 45.7 44.6 44.2
42.5 41.3
20.0 40.1
35.1
32.2 31.2
0.0
C
Ja
P
lS
an
er
cu
ar
an
ol
os
hi
ra
om
en
ol
ni
ru
uy
ua
on
ic
ai
ex
m
om
le
iv
te
ar
ti
ag
si
gu
ta
am
ad
ad
an
al
ez
te
du
ic
ai
in
ia
l
d
ag
va
ua
ay
m
ca
ue
or
a
bi
ic
R
ra
ic
al
ua
do
a
an
y
s
ta
la
a
a
r
te
R
s
ep
ub
lic
Source: AmericasBarometer2006 by LAPOP Error Bars: 95% CI
Figure XII-4. Support for the Institutions of the Rule of Law in Latin America and the Caribbean
What factors create popular support for the institutions that make up the rule of law?
Socio-demographic variables such as education, age and geographic location inside a country are
usually powerful explanatory variables for the generation of individual attitudes. For this study I
have designed a regression model which attempts to uncover the predictors for individual support
for the institutions of the rule of law in Latin America and the Caribbean. Table XI-2 contains
the result of the regression equation. It is surprising that variables such as the geographic location
of the citizens turn out not to be significantly related to the dependent variable. Institutions such
as the Police and lower courts usually lack a strong presence in rural settings, which causes
citizens in these areas to have a lower level of trust than citizens in urban areas. The difference,
however, turned out not to be significant. This does not mean, however, that this tendency does
not present itself in some countries; it just means that at the regional level, when one combines it
with other factors, geographic location inside a country turns out to be statistically insignificant
as a predictor of individual support for the institutions of the rule of law.
286
Table XII-2. Predictors for Individual Support of Institutions of the Rule of Law
Std.
B Error Sig.
(Constant) 44.251 8.950 0.000
Urban-Rural 3.049 2.545 0.232
Satisfaction with democracy 2.217 0.629 0.000
Corruption Victimization -0.055 0.025 0.028
Corruption in Courts -0.262 3.220 0.935
Perception of corruption -0.114 0.042 0.007
Crime Victimization -0.055 0.035 0.028
Education 0.121 0.310 0.696
Wealth measured by material possesions -0.021 0.724 0.977
Income -0.000 0.000 0.515
Mexico 0.833 8.004 0.917
Guatemala -2.894 6.765 0.669
Salvador -0.901 6.199 0.884
Honduras -17.65 8.789 0.048
Nicaragua 2.733 6.855 0.690
Panama 1.104 16.12 0.945
Colombia 8.657 17.49 0.621
Ecuador -19.19 7.115 0.007
Peru -8.721 6.666 0.128
Chile 0.935 15.87 0.975
Dominican Republic -12.49 6.000 0.038
Haiti 0.608 5.994 0.919
Jamaica -4.008 5.734 0.485
Guyana 2.679 5.436 0.622
Uruguay 15.99 14.31 0.264
Venezuela 18.31 15.85 0.248
R Squared 0.177
Other variables which do not have a significant effect of the levels of individual trust for
the institutions of the rule of law are economic factors, such as income and wealth measured in
ownership of material goods, such as household appliances, cell phones, functioning bathrooms
inside the house, etc. The data show that these variables have a positive but not significant effect
on individual levels of confidence in legal institutions.
Education turned out to be another insignificant factor in predicting individual support for
the institutions of the rule of law. Prior studies conducted by the Latin American Public Opinion
Project have shown that the relationship between education and institutional legitimacy is usually
negative. This means that those people with higher levels of education usually present lower
levels of institutional trust than those that are less educated. A correlation analysis between the
education and our index of individual trust for the institutions of the rule of law indeed shows
that when these two variables are the only ones included in the mix, the relationship is negative.
However, when one adds more ingredients to it, the effect of education is attenuated by personal
experiences, especially with corruption and crime.
287
Corruption is a problem that has plagued all Latin American democracies since their
rebirth in the late 70s and early 80s. In the last decade, several elected and appointed officials in
Latin America, including Presidents, as in the cases of Mexico, Costa Rica and Ecuador, have
faced criminal charges for acts of corruption committed during the exercise of their public
mandate. Corruption can be defined as the utilization of public goods for private benefits
(Weyland 1998) and it is a disease that has quickly proliferated and attracted the attention of the
academic and political communities as well as the International Organizations. Due to its nature,
corruption is extremely hard to measure (Buscaglia 2001, Seligson 2006) and that is why
academics have used various strategies to do so. One of the more widely used measures is
Transparency Internationals Corruption Perception Index, which is based in surveys carried out
to businessmen in different countries all over the world. In this chapter I will use two different
measures of corruption. One is similar to CPI and is based on a question from the questionnaire
of the AmericasBarometer.
EXC7. Keeping in mind your experience and what youve heard, is corruption among public officials:
(1) Completely generalized (2) Somewhat extensive
(3) A little extensive (4) Not generalized at all
The second way to measure corruption has is to capture the experiences of people with it.
Citizens in Latin America and the Caribbean are often victims of corruption by public officials,
such as police officers and judges and also by private entities, such as schools and even at the
workplace. The AmericasBarometer has developed a battery of questions that is aimed at
measuring the respondents experiences with corruption at various settings, like the courts, public
hospitals, schools and other public offices. Based on those questions we have created an index of
corruption victimization which measures the frequency with which our respondents were
victimized by corruption on the year prior to them answering our survey. For analytical
purposes, I have separated the question that measures corruption in the judiciary and inserted it
into the regression model separately, because of the relevance that corruption in the judicial
system should have on the generation of individual attitudes towards the institutions of the rule of
law. What the regression analysis shows is that not only those who have been victimized by
corruption in the courts, but those people who had to pay bribes in different settings have much
lower levels of trust than those who did not have to pay a bribe in the year prior to our survey, as
shown in Figure XII-5.
288
50.0
40.0
Law
30.0
20.0
10.0
0 1 2 3 4 5 6 7
Number of Ways a Person was Victimized by Corruption Last

Year
Figure XII-5. Corruption Victimization and Individual Trust for the Institutions of the Rule of Law
The respondents perception of corruption also turned out to be a determinant factor in

their support for the institutions that make out the rule of law. Just like with victimization, those
people who perceive a high level of corruption in the public sector show lower levels of
institutional support than their counterpart, which believe that corruption in the public sector is
not at all generalized. Figure XII-6 illustrates the results of the relationship between these two
variables. Two things should jump to the readers attention. First, the difference between the
groups with high and low levels of support for the institutions of the rule of law is not as drastic
as when we asked them about their personal experiences with corruption. Second, apparently
people who consider that corruption is not generalized at all in the public sector have lower levels
of confidence in the institutions of the rule of law than those who perceive that it is somewhat
generalized. The error bars at the top of the bars, however, show that the difference between one
and the other is not statistically significant, but just a product of the low number of people who
chose that option. Figure XII-6, then, brings us to the conclusion that the best way to measure
289
corruption is to look at it through the citizens personal experiences with it , because the results
this measure produces are more reliable.
60.0
50.0
40.0
Law
30.0
52.9
50.3 49.7
20.0 41.8
10.0
0.0
Generalized Somewhat A bit Generalized Not Generalized
Generalized
Perception of Corruption Among Public Officials

Figure XII-6. Perception of Corruption and Confidence in the Institutions of the Rule of Law
Personal experiences with crime also have an important effect in personal levels of
institutional support. For the rule of law to be in place, effective democratic institutions are
necessary but not sufficient. The rule of law also needs a collective understanding that the regime
type and the institutions are better than the ones in place before the transition to democracy, even
with all the problems faced by the new regimes (Cruz 2005). Violence and crime in Central
American countries have generated exclusion and diminished social interaction as well as trust in
the institutions of the rule of law in that part of the continent (Cruz 2004, Wielandt 2005). The
same is happening in the southern hemisphere, where crime victimization has been raising
steadily in the last decade. The AmericasBarometer questionnaire addresses the issue of crime in
a battery of questions, among which we have selected one to analyze in the present chapter.
290
VIC1. Have you been the victim of a crime in the last 12 months?
(1) Yes (2) No (8) NS/NR
Figure XII-7 ratifies the results presented in the linear regression model. Those
respondents who were victimized by a criminal act at least once in the year prior to the survey
show significantly lower levels of institutional trust than those who were not victims of a violent
crime in the last year.
50.0
40.0
30.0
Law
47.0
20.0 41.3
10.0
0.0
Yes No
Victimized by Crime in the Last 12 Months
Figure XII-7. Crime Victimization and Confidence in the Institutions of the Rule of Law
Throughout this section I have enumerated some of the factors that contribute to explain
the attitudes of citizens towards the institutions of the rule of law in their countries. With the
exception of education, most of the socio-demographic variables did not produce a significant
impact in the dependent variable. Neither family income and personal wealth measured in the
ownership of material goods nor geographical location inside a country have a statistically
291
significant effect individual attitudes towards the institutions that make up the rule of law. The
experiences of citizens with corruption and crime also play an important role in the formation of
individual attitudes towards the legal institutions, as is displayed in previous pages. There is one
part of the answer missing, however, and that part has to do with context. The regression model
presented in this section uses dichotomous variables for each country, with the objective of
analyzing the impact of context on individual support for the institutions for the rule of law,
taking Costa Rica as a reference. The results indicate that while the effect is not significant for
every country, it is in some of them, such as Ecuador, Peru, Chile and the Dominican Republic.
What does this means? How is it possible that a variable with the name of a country can
significantly affect the formation of individual attitudes towards the rule of law? The answer is
simple. What the results reveal is that the context in which one lives is important. For the
regression model presented in this section, dichotomous variables were devised for every country
in the sample except for Costa Rica. This is done so that the results can be compared to the
country that is omitted from the model. Lets take the example of Ecuador. The regression
coefficient for Ecuador is negative and statistically significant. This means that, ceteris paribus,
the average Ecuadorian reports a lower level of institutional trust than the average Costa Rican.
In simple terms, living in Ecuador has a negative impact on individual support for the institutions
that make up the rule of law.
This can be explained in a simple way. Ecuadors political environment in the last decade
has been characterized by instability and judicial insecurity, especially due to the events that have
taken place the last couple years with the judicial sector (Incln 2006, Meja Acosta, et al. 2006).
At the end of 2004, the legislature, with a simple majority, came up with a resolution that
expelled 27 of the 31 Supreme Court Justices. The reason given by Congress was that through
that resolution they were purging a court that only certain political interest groups. However,
Congress made the mistake of building a new Supreme Court packed with judges that were
linked with all of the political parties which made up the congressional majority. This decision
was rejected by the majority of the population; and intense popular pressure forced then President
Lucio Gutierrez to disqualify the new Supreme Court Justices. The mechanism to elect a new
Supreme Court was never devised, since Gutierrez was overthrown five days later. The judicial
system was shut down for a year, since its main institution did not function throughout the
duration of the selection process of the 31 new justices. Although the new judges were selected
under heavy scrutiny from national and international organizations, it has failed to obtain popular
trust and support (Donoso, et al. 2006). The situation has only gotten worse by a series of
irregularities committed by the National Chief of Police. Four Chiefs of Police were replaced in
only two weeks due to incompetence and allegations of corruption. These events, compared to
the relative political stability that has reigned in Costa Rica, exemplify the importance of context
in the generation of individual attitudes in Latin America and the Caribbean.
292
Democracy and the Rule of Law
In the section leading up to this one, the discussion focused on the structural and
individual predictors of citizen support for the institutions of the rule of law in Latin America and
the Caribbean. We know that variables such as the strength of the judicial system, the experience
of our respondents with crime and corruption, the level of education and the socio-political
context are important in the generation of individual attitude towards the institutions of the rule
of law. But is this relevant for the democratic consolidation of the region? The answer is
definitely yes. The last few paragraphs of this chapter will be dedicated to demonstrate the
importance of citizen support for the institutions of the rule of law in the survival of
representative democracy.
The third wave of democratization begun in Latin America over 25 years ago and still
democratic consolidation continues to be perhaps the greatest challenge of the regions nations.
The many flaws in the administration of justice systems in the continent and the lack of security
faced by citizens play a key role in this problem(Binder 1991, Buscaglia, et al. 2000). To better
illustrate the relationship between popular support for the institutions of the rule of law and
democratic consolidation I will again turn to the data gathered by the 2006-2007
AmericasBarometer. There are many ways to measure support for democracy in a country. One
of them is to see if the general population thinks of democracy as the only game in town. This
means that they prefer democracy to any other form of government, even with the problems that
come with a democratic regime. The question formulated in our questionnaire is:
ING4. Democracy may have its problems but it is still better than any other form of government. To what extent do
you agree with this phrase?
The respondents were shown a scale from 1 to 7 where one equals fully disagree and
seven represents fully agree. Figure XII-8 shows the effect that support for the institutions of
the rule of law has on the legitimacy of democracy as a form of government. The relationship
between the two variables is as expected, with those reporting high levels of institutional trust
also presenting high levels of acceptance for democracy as the best form of government. The
graph also gives us some encouraging news. Even those people with extremely low levels of
trust for the institutions that make up the rule of law are in the higher end of the scale of
democratic desirability. This means that even citizens who completely distrust legal institutions
believe in democracy. But what kind of democracy do they want?
293
55.0
50.0
45.0
Law
40.0
35.0
30.0
Completely 2 3 4 5 6 Completely
Disagree Agree
Democracy is the Best Form of Government
Figure XII-8. Support for the institutions of the rule of law and acceptance of democracy as the best
form of government
In the last few years, observers and academics have been witnesses to the rise of what
some now call hybrid democracies in the region. These democracies hold free and fair
elections, but ignore some of the basic principles of liberal and representative democracies, such
as the separation of powers, the right to private property and the guarantee of civil liberties.
While many people question these kinds of democracies, some of these restrictions are supported
by some people, who, tired of unfulfilled promises, are willing to give something up in exchange
for some stability and personal economic growth.
A lot has been said in recent years about the rebirth of populism in Latin America.
Populist regimes are characterized by strong and charismatic leaders who tend to woo the crowds
by generating in them a feeling of identity. People who vote for populist leaders feel that they
understand their needs and problems better than anyone and that they will work tirelessly to solve
294
them. This includes dodging legal obstacles to get things done faster. The legal systems of
many Latin American and Caribbean countries are slow and inefficient. Many countries in the
region still have not resorted to oral trials, which makes legal processes long and tedious. Figure
XII-9 illustrates the result of asking people about the usefulness of judges in a democratic regime.
POP3.
1.Judges frequently get in the way of our presidents and they should be ignored.
2. Even when judges sometimos impede the labor of our presidents, their decisions must always be respected.
As the graph shows, there is a significant difference both statistically and substantively
between those who believe that judges are a necessity and those who think they are not
indispensable, with regards to their levels of trust in the institutions of the rule of law.
50.0
Mean Trust in the Institutions of the Rule of Law
40.0
30.0
47.1
43.6
20.0
10.0
0.0
Judges are an Obstacle to the President Judges are Necessary Even if
and Should be Bypassed Sometimes they are Inconvenient
Figure XII-9. Support for the institutions of the rule of law and perception of the necessity of judges
295
Finally, there are people willing to justify extreme measures to improve the
administration of justice in their countries. The Supreme Court is not the most trusted institution
in Latin America and the Caribbean. In a scale from 0 to 100, the mean of support for this
institution barely breaks 40. This makes the Supreme Court one of the least trusted institutions in
the region, outperforming only Congress and the political parties. A battery of questions
designed by the AmericasBarometer attempts to capture the peoples tolerance for authoritarian
and antidemocratic measures, such as shutting down a countrys Supreme Court.
JC16. Do you think that there could ever be a reason to justify the closing of the Supreme Court by the President?
(1) Yes (2) No
Figure XII-10 shows that people who would justify an eventual closing of the Supreme
Court by the president have lower levels of trust for the institutions of the rule of law that their
counterparts.
50.0
40.0
30.0
Law
47.3
42.4
20.0
10.0
0.0
Yes No
Is it Justifiable for the President to Shut Down the Supreme

Court?
Figure XII-10. Trust in the institutions of the rule of law and justification for the closing of the Supreme
Court
296
Conclusion
This chapter has attempted to uncover the factors that determine individual support for the
institutions of the rule of law in Latin America and the Caribbean. I started by underlining the
importance of structural variables, such as the need for a strong judiciary, capable of ruling
against the government without fear of repercussions. I also reviewed some of the individual
predictors of institutional support, from which corruption and crime victimization as well as
context stand out. Finally, I observed the dangerous consequences that the lack of trust in the
institutions of the rule of law can bring to liberal democracy in the region. Citizens who are
disappointed with the way their governments provide justice and security are willing to tolerate
drastic and authoritarian measures in order to fell safer. It is important that the governments in
the continent work to control the raising levels of crime and corruption and to build solid and
autonomous institutions with efficient mechanisms of accountability that can generate trust and
support.
297
References
Becker, David. "Latin America: Beyond Democratic Consolidation." Journal of Democracy 10,
no. 2 (1999): 138-51.
Binder, Alberto. "Reform of the Penal System in Latin America." Arlington, VA: Council of
State Courts, 1991.
Buscaglia, Edgardo. "Judicial Corruption in Developing Countries: Its Causes and Economic
Consequences." In Essays in Public Policy. Washington, DC: Hoover Institute, 2001.
Buscaglia, Edgardo, Valeria Merino, and Ana Lucia Jaramillo. "Estudio Sobre La Correlacion
Entre La Existencia De Justicia Y La Consolidacion De La Democracia En Ecuador. Quito:
Corporacion Latinoamericana para el Desarrollo, 2000.
Cappelletti, Mauro. "Who Watches the Watchmen? A Comparative Study on Judicial
Responsability." In Judicial Independence: The Contemporary Debate, edited by S.
Shetreet and J. Deschenes. Dordrecht: martinus Nijhoff Publishers, 1985.
Carothers, Thomas. Promoting the Rule of Law Abroad: In Search of Knowledge. Washington
D.C.: Carnegie Endowment for International Peace, 2006.
Collier, David, Fernando Hidalgo, and Andra Maciuceanu. "Essentially Contested Concepts:
Debates and Applications." Journal of Political Ideolofies 11, no. 3 (2006): 211-46.
Cross, Frank B. "The Relevance of Law in Human Rights Protection." International Review of
Law and Economics 19 (1999): 87-98.
Cruz, Jose Miguel. "Pandillas Y Capital Social En Centroamerica." In Maras Y Pandillas De
Centroamerica: UCA Editores, 2004.
. "Violencia Y Democratizacion En Centro America: El Impacto Del Crimen En Los
Regimenes De La Post Guerra." PNUD, 2005.
Dakolias, Maria. "Ecuador: Judicial Sector Assesment." Washington D.C: The World Bank,
1994.
Donoso, Juan Carlos, Daniel Moreno, Diana Maria Orces, Vivian Schwartz-Blum, and Mitchell
A. Seligson. "Democracy Audit: Ecuador 2006." Quito: Latin American Public Opinion
Project, 2006.
Fiss, Owen. "The Limits of Judicial Independence." University of Miami Inter-American Law
Review 25 (1993).
Goertz, Gary. Social Science Concepts: A User's Guide: Princeton University Press, 2006.
Holston, James, and Teresa P.R. Caldeira. "Democracy, Law and Violence: Disjunctions of
Brazilian Citizenship." In Fault Lines of Democracy in Post-Transition Latin America,
edited by Felipe Aguero. Coral Gables: North-South Center Press, 1998.
Howard, Robert, and Henry Carey. "Is an Independent Judiciary Necessary for Democracy?"
Judicature 87, no. 6 (2004): 284-91.
Huntington, Samuel. The Third Wave: Democratization in the Late Twentieth Century
Norman: University of Oklahoma Press, 1991.
Incln, Silvia. "Fragile Judicial Independence in Latin America: A Decade after Judicial
Reform." In Annual Meeting of the Midwest Political Science Association. Chigago, 2006.
Larkins, Christopher. "Judicial Independence and Democratization: A Theoretical and
Conceptual Analysis." The American Journal of Comparative Law 44, no. 4 (1996): 605-
26.
298
Meja Acosta, Andres, Maria Caridad Araujo, Anibal Perez-Linan, and Sebastian Saiegh. "Veto
Players, Fickle Institutions and Low Quality Policies: The Policy Making Process in
Ecuador (1979-2005)." Inter-American Development Bank, 2006.
O'Donell, Guillermo. "Delegative Democracy." Journal of Democracy 5 (1994): 55-69.
O'Donnell, Guillermo, and Philippe Schmitter. Transitions from Authoritarian Rule: Tentative
Conclusions About Uncertain Democracies. Baltimore: John Hopkins University Press,
1986.
Pereira, Anthony. "An Ugly Democracy? State Violence and the Rule of Law in Post-
Authoritarian Brazil." In Democratic Brazil: Actors, Institutions and Processes., edited by
Peter Kingstone Jr and Thimoty J. Powers. Pittsburgh: University of Pittsburgh's Press,
2000.
Prillaman, William. The Judiciary and Democratic Decay in Latin America: Declining
Confidence in the Rule of Law. Westport: Praeger, 2000.
Rios-Figueroa, Julio. "Judicial Independence: Definition, Measurement and Its Effects on
Corruption: An Analysis of Latin America." Doctoral Dissertation, New York University,
2006.
Schedler, Andreas, Larry Diamond, and Marc F. Plattner. The Self Restraining State: Power and
Accountability in New Democracies. Boulder: Lynne Riener, 1999.
Schor, Miguel. "The Rule of Law and Democratic Consolidation in Latin America." In Latin
American Studies Association. Dallas, Texas, 2003.
Seligson, Mitchell. "Auditoria De La Democracia: Ecuador 2004." edited by CEDATOS.
Quito, 2004.
. "The Measurement and Impact of Corruptiuon Victimization: Survey Evidence in Latin
America." World Development 34, no. 2 (2006): 381-404.
Seligson, Mitchell, and Edward N. Muller. "Civic Culture and Democracy: The Question of the
Causal Relationship." American Political Science Review 88 (1994): 635-54.
Tate, Neal, Linda Camp Keith, and Steven Poe. "Repression of the Human Right to Personal
Integrity Revisited: A Global Cross-National Study Covering the Years 1976-1993."
International Studies Quarterly 43 (1999): 291-315.
Verner, Joel G. "The Independence of Supreme Courts in Latin America: A Review of the
Literature." Journal of Latin American Studies 16, no. 2 (1984): 463-506.
Weyland, Kurt. "The Politics of Corruption in Latin America." Journal of Democracy 9, no. 2
(1998): 108-21.
Wielandt, Gonzalo. "Hacia La Construccion De Lecciones Del Posconflicto En America Latina
Y El Caribe. Una Mirada a La Violencia Juvenil En Centroamerica." Santiago de Chile:
CEPAL, 2005.
www.humanrightsdata.org. "The Cingranelli and Richards Human Rights Dataset." (2004).
www.state.gov. U.S. State Department's Annual Human Rights Country Reports 2005 [cited].
Yamanashi, David Scott. "Judicial Independence and Human Rights." In American Political
Science Association Annual Meeting, 2002.
299

Challenges To Democracy in Latin America

Enviado por

Dados do documento

Título original

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Challenges To Democracy in Latin America

Enviado por

Direitos autorais:

Formatos disponíveis

Chhaalllleennggeess ttoo Deemmooccrraaccyy iinn

Laattiinn Ammeerriiccaa aanndd tthhee Caarriibbbbeeaann::

Mitchell A. Seligson, Ph.D., Editor

II. Illiberal Democracy and Normative Democracy: How is Democracy Defined in

V. Decentralize or Centralize? Challenges for Reform of the State and Democracy

Part II. Challenges to Democracy from Civil Society................................................. 145

References ............................................................................................................................................. 171

VIII. National Identity and Ethnic Minorities in the Americas................................. 177

IX. Challenges of Tolerance in the Americas ............................................................. 195

Part III. Challenges to the Rule of Law ....................................................................... 217

The Level of Corruption in the Region ................................................................................................. 253

Promoting democracy and good governance is a US government foreign policy priority,

Dr. Margaret Sarles

I am very pleased to introduce to you the 2006-2007 round of the AmericasBarometer

An additional technological innovation in the 2006-2007 round is that we used handheld

There is much to celebrate in the strengthening of democratic practices and institutions in

This research is based on the AmericasBarometer democracy surveys undertaken in 22

We have chosen to focus on a single, important theme: the challenges to democracy in

Many researchers have seen this political/cultural characteristic as an indication of the

Carrins approach is methodologically path-breaking, particularly in the Latin American

Like other contributors to this volume, Boidi is struck by what appear to be

Schwarz-Blum reiterates an important initial point: In highly industrialized countries,

Moreno introduces the concept of a sense of national integration, a combination of pride

There is a complex relationship between victimization, perceptions of insecurity, and

I. The Rise of Populism and the Left: Challenge to Democratic

The Political Orientation of Latin Americans is Important to Democracy

Political Orientation Matters: People Vote their Beliefs

National ideology mean = 5.9

Oscar Arias (PLN)

Oscar Arias (PLN)

Ricardo Jaime Toledo (PUSC)

Ricardo Jaime Toledo (PUSC)

Otton Sols (PAC)

Otton Sols (PAC)

Otto Guevara (ML)

Otto Guevara (ML)

Left Ideology Right

Horizontal lines indicate 95% confidence interval

D aniel Ortega, FSLN

S chaf ik H nda l (FM LN)

Anton io S aca ( AREN A)

Antonio Saca (A RENA)

Ho riz on tal lin es i nd ica te 95 % co nfide nc e int erva ls

Left Ideology Right

Leftists are Less Likely to Favor some Core Democracy Values

We used a number of methods. First, we developed a scale based on a question often

However, by another standard measure of democratic values in the surveys, political

Left Ideology Right

Latin Americans are not Ready to Jettison Democratic Institutions

None Primary Secondary University

Conclusions: Political Institutions and Values Remain Important

Taggart, Paul A. Populism, Concepts in the Social Sciences. Buckingham [England] ;

II. Illiberal Democracy and Normative Democracy: How is

Mass Conceptualizations of Democracy

The 2006-2007 AmericasBarometer questionnaire asked the following open-ended

Table II-1. Mass Conceptualizations of Democracy, 2006-2007

Frequency Percent Valid Percent

96% Can't define

Figure II-2. Support for Democracy by the Ability to Define it

second case, it is instrumental because it is based on a means/end rationality. Democracy can

Table II-2. Conceptualizations of Democracy Based on the Rationality of the Belief

Figure II-3. Conceptualizations of and Support for Democracy

How Consistent is the Support for Democracy?